Merge pull request 'feat(p10-1C-JK): Java + Kotlin AST chunkers — JVM family 코드 색인 활성화' (#152) from feat/p10-1c-jk into main

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Cargo.lock

@@ -4127,7 +4127,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-app"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "base64 0.22.1",
@@ -4172,7 +4172,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-chunk"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
@@ -4187,7 +4187,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-cli"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "clap",
@@ -4208,7 +4208,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-config"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "dirs 5.0.1",
@@ -4223,7 +4223,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-core"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
@@ -4237,7 +4237,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-embed"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
@@ -4251,7 +4251,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-embed-local"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "fastembed",
@@ -4264,7 +4264,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-eval"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "kebab-app",
@@ -4283,7 +4283,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-llm"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "kebab-core",
@@ -4292,7 +4292,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-llm-local"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "kebab-config",
@@ -4309,7 +4309,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-mcp"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "kebab-app",
@@ -4327,7 +4327,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-normalize"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "kebab-core",
@@ -4342,7 +4342,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-parse-code"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "gix",
@@ -4352,12 +4352,18 @@ dependencies = [
  "time",
  "tracing",
  "tree-sitter",
+ "tree-sitter-go",
+ "tree-sitter-java",
+ "tree-sitter-javascript",
+ "tree-sitter-kotlin-ng",
+ "tree-sitter-python",
  "tree-sitter-rust",
+ "tree-sitter-typescript",
 ]
 
 [[package]]
 name = "kebab-parse-image"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "ab_glyph",
  "anyhow",
@@ -4381,7 +4387,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-parse-md"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "kebab-core",
@@ -4398,7 +4404,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-parse-pdf"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
@@ -4411,7 +4417,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-parse-types"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "kebab-core",
  "serde",
@@ -4419,7 +4425,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-rag"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
@@ -4440,7 +4446,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-search"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "globset",
@@ -4459,10 +4465,11 @@ dependencies = [
 
 [[package]]
 name = "kebab-source-fs"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
+ "globset",
  "ignore",
  "kebab-config",
  "kebab-core",
@@ -4477,7 +4484,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-store-sqlite"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "blake3",
@@ -4498,7 +4505,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-store-vector"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "arrow",
@@ -4522,7 +4529,7 @@ dependencies = [
 
 [[package]]
 name = "kebab-tui"
-version = "0.7.0"
+version = "0.13.0"
 dependencies = [
  "anyhow",
  "crossterm",
@@ -8523,12 +8530,62 @@ dependencies = [
  "tree-sitter-language",
 ]
 
+[[package]]
+name = "tree-sitter-go"
+version = "0.25.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c8560a4d2f835cc0d4d2c2e03cbd0dde2f6114b43bc491164238d333e28b16ea"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
+[[package]]
+name = "tree-sitter-java"
+version = "0.23.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0aa6cbcdc8c679b214e616fd3300da67da0e492e066df01bcf5a5921a71e90d6"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
+[[package]]
+name = "tree-sitter-javascript"
+version = "0.25.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68204f2abc0627a90bdf06e605f5c470aa26fdcb2081ea553a04bdad756693f5"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
+[[package]]
+name = "tree-sitter-kotlin-ng"
+version = "1.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e800ebbda938acfbf224f4d2c34947a31994b1295ee6e819b65226c7b51b4450"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
 [[package]]
 name = "tree-sitter-language"
 version = "0.1.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "009994f150cc0cd50ff54917d5bc8bffe8cad10ca10d81c34da2ec421ae61782"
 
+[[package]]
+name = "tree-sitter-python"
+version = "0.25.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6bf85fd39652e740bf60f46f4cda9492c3a9ad75880575bf14960f775cb74a1c"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
 [[package]]
 name = "tree-sitter-rust"
 version = "0.24.2"
@@ -8539,6 +8596,16 @@ dependencies = [
  "tree-sitter-language",
 ]
 
+[[package]]
+name = "tree-sitter-typescript"
+version = "0.23.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6c5f76ed8d947a75cc446d5fccd8b602ebf0cde64ccf2ffa434d873d7a575eff"
+dependencies = [
+ "cc",
+ "tree-sitter-language",
+]
+
 [[package]]
 name = "try-lock"
 version = "0.2.5"

Cargo.toml

@@ -31,7 +31,7 @@ edition       = "2024"
 rust-version  = "1.85"
 license       = "MIT OR Apache-2.0"
 repository    = "https://github.com/altair823/kebab"
-version       = "0.7.0"
+version       = "0.13.0"
 
 [workspace.dependencies]
 anyhow       = "1"
@@ -90,6 +90,15 @@ gix          = { version = "0.70", default-features = false, features = ["revisi
 # chunker stays tree-sitter-free — AST work is parser-side per design §6.3.
 tree-sitter      = "0.26"
 tree-sitter-rust = "0.24"
+# Python / TS / JS grammars for code ingest (kebab-parse-code, p10-1B).
+tree-sitter-python     = "0.25.0"
+tree-sitter-typescript = "0.23.2"
+tree-sitter-javascript = "0.25.0"
+# Go grammar for code ingest (kebab-parse-code, p10-1C-Go).
+tree-sitter-go         = "0.25.0"
+# JVM family grammars for code ingest (kebab-parse-code, p10-1C-JK).
+tree-sitter-java       = "0.23.5"
+tree-sitter-kotlin-ng  = "1.1.0"   # bare tree-sitter-kotlin requires ts <0.23; -ng uses tree-sitter-language 0.1 (ts 0.26 compat)
 
 # Disk-footprint trim for dev / test builds. Codegen, opt-level, and
 # behavior are unchanged — only DWARF debug info is reduced (line

HANDOFF.md

@@ -4,7 +4,7 @@
 
 ## 한 줄 요약
 
-P0–P5 + P6 + P7 + P9-1/2/3/4 (Library / Search / Ask / Inspect) 머지 완료. `kebab ingest` 가 markdown / image / PDF 모두 처리. `kebab search` / `kebab ask` 가 매체 가로질러 결과 + page citation 반환. `kebab tui` 가 4 패널 (Library + Search + Ask + Inspect) 제공 — 사용자가 `?` 로 ask, `/` 로 search, Library Enter / Search `i` 로 inspect, Search `g` 로 editor jump. 다음 후보 = P9-5 (desktop tauri) 또는 보류 중인 P8 (audio) 의 시스템 dep brainstorm.
+P0–P5 + P6 + P7 + P9-1/2/3/4 (Library / Search / Ask / Inspect) 머지 완료. `kebab ingest` 가 markdown / image / PDF / 소스코드 (Rust / Python / TS / JS / Go / Java / Kotlin) 처리. `kebab search` / `kebab ask` 가 매체 가로질러 결과 + page / code citation 반환. `kebab tui` 가 4 패널 (Library + Search + Ask + Inspect) 제공. P10-1C (Go + Java + Kotlin) 완료 — 다음 후보 = P10-1D (C/C++) 또는 P9-5 (desktop tauri) 또는 보류 중인 P8 (audio).
 
 ## Phase 로드맵
 
@@ -20,7 +20,7 @@ P0–P5 + P6 + P7 + P9-1/2/3/4 (Library / Search / Ask / Inspect) 머지 완료.
 | **P7** | PDF text + page citation | `kebab-parse-pdf` | P5 | ✅ 완료 (3/3 component, page-level chunker + ingest wiring) |
 | **P8** | 음성 transcription + timestamp citation | `kebab-parse-audio` | P5 | ⏸ 보류 (whisper-rs 시스템 dep brainstorm 필요) |
 | **P9** | TUI + desktop app | `kebab-tui`, `kebab-desktop` | P5 | 🟡 진행 (4/5 component — P9-1/2/3/4 완료 [Library / Search / Ask / Inspect], P9-5 desktop 예정 · 도그푸딩 피드백 **20/20 ✅**) |
-| **P10** | code ingest framework | `kebab-parse-code` | P5 | 🟡 진행 중 — 1A-1 ✅ (wire schema + parse-code skeleton + filter flags), 1A-2 ✅ (Rust AST chunker, tree-sitter-rust, `code-rust-ast-v1` — kebab 자기 dogfooding 가능, v0.7.0) |
+| **P10** | code ingest framework | `kebab-parse-code` | P5 | 🟡 진행 중 — 1A-1 ✅ (wire schema + parse-code skeleton + filter flags), 1A-2 ✅ (Rust AST chunker, `code-rust-ast-v1` — v0.7.0), 1B ✅ (Python/TS/JS AST chunkers — v0.8.0 이후), **1C-Go ✅ (Go AST chunker, `code-go-ast-v1` — v0.12.0)**, **1C-JavaKotlin ✅ (Java + Kotlin AST chunkers, `code-java-ast-v1` / `code-kotlin-ast-v1` — v0.13.0)** |
 
 P0~P5 직렬. P6~P9 P5 이후 병렬 가능.
 
@@ -32,6 +32,7 @@ P0~P5 직렬. P6~P9 P5 이후 병렬 가능.
 
 머지 후 발견된 모든 deviation / hotfix 의 dated 로그는 [tasks/HOTFIXES.md](tasks/HOTFIXES.md). 본 요약은 \"누군가가 인수받을 때 알아두면 시간을 많이 절약하는\" 항목만:
 
+- **2026-05-20 P10-1B (Rust 1A symbol path 비일관 + expression-level 함수 미방출)** — (a) Rust `code-rust-ast-v1` 은 file-scope nesting 만 (workspace path prefix 없음), 1B 의 Python/TypeScript/JavaScript 는 workspace 경로 → module path prefix 사용 (비일관 수용, retrofit = chunker_version bump + reindex 필요, 사용자 명시 요청까지 보류); (b) TS/JS 의 `const foo = () => {...}` 같은 expression-level 함수는 `<top-level>` glue 로 처리됨 (declaration-level 단위만 1B 1차 범위). 자세한 내용: `tasks/HOTFIXES.md` (2026-05-20) 두 항목.
 - **2026-05-19 P10-1A-2 (code_rust_ast_v1.rs + SourceType)** — `AST_CHUNK_MAX_LINES` 상수가 `IngestCodeCfg.ast_chunk_max_lines` 를 읽지 않고 모듈 상수 200 고정 (Chunker trait 이 per-medium config 미노출); `SourceType::Code` variant 부재로 code 파일이 `SourceType::Note` 로 분류됨 — 두 항목 모두 `tasks/HOTFIXES.md` (2026-05-19) 에 기록.
 - **2026-05-07 fb-26 (progress.rs)** — `Aborted` unconditional writeln (TTY duplicate) + `Completed` TTY no summary fixed; `KEBAB_PROGRESS=plain` env + quiet suppression added
 - **2026-05-07 fb-28 (main.rs)** — `--readonly` (KEBAB_READONLY) blocks Ingest/IngestFile/IngestStdin/Reset; `--quiet` suppresses progress stderr; error.v1 code: "readonly_mode"

README.md

@@ -34,7 +34,7 @@ cargo install --git https://gitea.altair823.xyz/altair823-org/kebab.git --bin ke
 
 업데이트는 `git pull && cargo install --path crates/kebab-cli --locked --force` 또는 git URL 형식의 경우 `cargo install --git ... --force`.
 
-제거는 `cargo uninstall kebab-cli`. 이 명령은 binary 만 지우고 워크스페이스 데이터는 그대로 남는다. 데이터까지 정리하려면 `kebab reset --all --yes` (config + data + cache + state 4 개 XDG 경로 모두 wipe — **irreversible**, 재시작 시 `kebab init` 다시 실행). 부분 wipe 는 `kebab reset --data-only` (config 보존), `kebab reset --vector-only` (Lance + `embedding_records` 만, 다음 ingest 가 re-embed) 등.
+제거는 `cargo uninstall kebab-cli`. 이 명령은 binary 만 지우고 워크스페이스 데이터는 그대로 남는다. 데이터까지 정리하려면 `kebab reset --all --yes` (config + data + cache + state 4 개 XDG 경로 모두 wipe — **irreversible**, 재시작 시 `kebab init` 다시 실행). 부분 wipe 는 `kebab reset --data-only` (config 보존), `kebab reset --vector-only` (Lance + `embedding_records` 만, 다음 ingest 가 re-embed), **`kebab reset --orphans-only`** (현재 walker scope 밖에 있는 stored doc 만 정리 — `config.workspace.include` 좁히거나 sub-dir 옮긴 후 explicit reconcile; fs 의 file 은 건드리지 않음) 등.
 
 ## Quick start
 
@@ -42,7 +42,7 @@ cargo install --git https://gitea.altair823.xyz/altair823-org/kebab.git --bin ke
 # 첫 실행 — XDG 경로에 데이터 디렉토리 + config.toml 생성
 kebab init
 
-# config 손보고 — workspace.root, 모델 endpoint 등 설정 (지원 형식: md / png / jpg / pdf / rs)
+# config 손보고 — workspace.root, 모델 endpoint 등 설정 (지원 형식: md / png / jpg / pdf / rs / py / ts / js / go)
 ${EDITOR:-vi} ~/.config/kebab/config.toml
 
 # 색인 (Markdown / 이미지 / PDF 모두 한 번에)
@@ -70,7 +70,7 @@ kebab doctor
 | 명령 | 동작 |
 |------|------|
 | `kebab init` | XDG 경로에 데이터 디렉토리 + config.toml 생성 |
-| `kebab ingest [<path>]` | Markdown / 이미지 / PDF / Rust 소스코드 색인 (idempotent). TTY 에서는 stderr 진행 바, non-TTY (CI / pipe) 는 stderr 한 줄씩, `--json` 은 stdout 에 `ingest_progress.v1` 라인 streaming 후 마지막에 `ingest_report.v1`. Ctrl-C 한 번이면 현재 asset 마무리 후 abort (부분 commit 보존, idempotent re-run), 두 번째 Ctrl-C 는 hard exit. Markdown title 이 frontmatter 에 없어도 첫 H1 → H2 → 첫 paragraph 80 자 → 파일명 순으로 자동 채움 (parser_version `md-frontmatter-v2`) — 기존 색인된 doc 도 다음 ingest 에서 새 title 로 갱신. **Incremental** (p9-fb-23): 두 번째 이후의 ingest 는 변하지 않은 doc (blake3 + parser/chunker/embedder version 모두 동일) 의 parse/chunk/embed/vector upsert 를 자동 스킵. final summary 에 `N unchanged` 카운트 표시. `--force-reingest` 로 skip 무시 강제 재처리. **지원 형식** (extractor 자동 결정 — config 에 명시 불가): Markdown (`.md`), 이미지 (`.png` / `.jpg` / `.jpeg`, OCR + caption), PDF (`.pdf`), **Rust 소스코드** (`.rs`, tree-sitter AST chunker `code-rust-ast-v1` — p10-1A-2). 다른 확장자는 자동 skip — `IngestItem.warnings` 에 사유 (`"unsupported media type: .docx"` 등), `IngestReport.skipped_by_extension` 에 카운트 분류, CLI / TUI summary 에 breakdown 표시. 코드 chunk 는 `citation.kind = "code"` 에 `citation.lang = "rust"` + `symbol` + line range 를 담고, SearchHit top-level 에 `code_lang = "rust"` + `repo` (`.git/` walk-up 의 디렉토리 이름) 가 backfill 됨. `--code-lang rust` / `--media code` filter 로 코드 전용 검색 가능 (p10-1A-1 filter flags). |
+| `kebab ingest [<path>]` | Markdown / 이미지 / PDF / Rust 소스코드 색인 (idempotent). TTY 에서는 stderr 진행 바, non-TTY (CI / pipe) 는 stderr 한 줄씩, `--json` 은 stdout 에 `ingest_progress.v1` 라인 streaming 후 마지막에 `ingest_report.v1`. Ctrl-C 한 번이면 현재 asset 마무리 후 abort (부분 commit 보존, idempotent re-run), 두 번째 Ctrl-C 는 hard exit. Markdown title 이 frontmatter 에 없어도 첫 H1 → H2 → 첫 paragraph 80 자 → 파일명 순으로 자동 채움 (parser_version `md-frontmatter-v2`) — 기존 색인된 doc 도 다음 ingest 에서 새 title 로 갱신. **Incremental** (p9-fb-23): 두 번째 이후의 ingest 는 변하지 않은 doc (blake3 + parser/chunker/embedder version 모두 동일) 의 parse/chunk/embed/vector upsert 를 자동 스킵. final summary 에 `N unchanged` 카운트 표시. `--force-reingest` 로 skip 무시 강제 재처리. **지원 형식** (extractor 자동 결정 — config 에 명시 불가): Markdown (`.md`), 이미지 (`.png` / `.jpg` / `.jpeg`, OCR + caption), PDF (`.pdf`), **소스코드** (`.rs` → `code-rust-ast-v1`, `.py` → `code-python-ast-v1`, `.ts`/`.tsx` → `code-ts-ast-v1`, `.js`/`.mjs`/`.cjs`/`.jsx` → `code-js-ast-v1`, `.go` → `code-go-ast-v1`, `.java` → `code-java-ast-v1`, `.kt`/`.kts` → `code-kotlin-ast-v1` — 모두 tree-sitter AST chunker). 다른 확장자는 자동 skip — `IngestItem.warnings` 에 사유 (`"unsupported media type: .docx"` 등), `IngestReport.skipped_by_extension` 에 카운트 분류, CLI / TUI summary 에 breakdown 표시. 코드 chunk 는 `citation.kind = "code"` 에 `citation.lang = "<lang>"` + `symbol` + line range 를 담고, SearchHit top-level 에 `code_lang` + `repo` (`.git/` walk-up 의 디렉토리 이름) 가 backfill 됨. `--code-lang rust` / `--code-lang python` / `--code-lang typescript` / `--code-lang javascript` / `--code-lang go` / `--code-lang java` / `--code-lang kotlin` / `--media code` filter 로 언어별·코드 전용 검색 가능 (p10-1A-1 filter flags). Python symbol 은 workspace 경로 → dotted module path prefix (예: `kebab_eval.metrics.compute_mrr`), TS/JS symbol 은 slash-style module path prefix (예: `src/Foo.Foo.search`), Go symbol 은 `package.Func` / `package.(*Receiver).Method` 형식, Java / Kotlin symbol 은 `com.foo.Foo.bar` 형식 (패키지 + 클래스 + 메서드/필드). |
 | `kebab search --mode {lexical,vector,hybrid} "<query>" [--no-cache] [--max-tokens N] [--snippet-chars N] [--cursor <opaque>] [--tag T] [--lang L] [--path-glob G] [--trust-min LEVEL] [--media TYPE] [--ingested-after RFC3339] [--doc-id ID] [--trace] [--bulk] [--repo NAME ...] [--code-lang LIST]` | 검색. hybrid는 RRF fusion, citation 포함. 같은 process 안에서 동일 query (NFKC + trim + lowercase 정규화) 반복 시 in-process LRU 캐시 hit (capacity = `[search] cache_capacity`, default 256). `--no-cache` 로 강제 bypass — 디버깅용. ingest commit 발생 시 `kv['corpus_revision']` bump 으로 모든 entry 자동 stale. **`--max-tokens` / `--snippet-chars` / `--cursor` (p9-fb-34)** — agent budget controls. `--json` 출력은 `search_response.v1` wrapper (`{hits, next_cursor, truncated}`) — pre-fb-34 의 bare array 와 호환 안 됨. mismatched cursor → `error.v1.code = stale_cursor`. **filter flags (p9-fb-36):** `--tag` 는 반복 가능 flag (`--tag rust --tag async`) 로 OR 매칭, `--media` 는 `,` 구분 다중 값 OR 매칭, 나머지 flags 간은 AND 조합. `--trust-min` 은 `primary\|secondary\|generated` 중 하나 (해당 level 이상 포함). `--ingested-after` 는 RFC3339 UTC — 파싱 실패 시 `error.v1.code = config_invalid` (exit 2). `--media md` 는 `markdown` alias 로 정규화. 알 수 없는 `--media` 값은 무조건 empty hits (오류 아님). **`--trace` (p9-fb-37)** — `search_response.v1.trace` 에 lexical / vector pre-fusion 후보 + RRF union + per-stage timing (`lexical_ms` / `vector_ms` / `fusion_ms` / `total_ms`) 노출. trace 요청은 캐시 우회 (`--no-cache` 없이도 항상 cold). **`--bulk` (p9-fb-42)** — stdin ndjson 으로 N query 한 번에 실행. `--json` 면 stdout per-query ndjson (`bulk_search_item.v1`) + stderr summary (`bulk_summary: total=N succeeded=S failed=F`). Cap 100. agent 가 query decomposition 후 sub-query 일괄 실행 시 single round-trip — App instance 재사용으로 캐시 / embedder cold-start 비용 한 번만. Per-query failure 는 item 의 `error` (error.v1) 에 격리, 다른 query 계속 진행. **code corpus filters (p10-1A-1):** `--repo` 는 반복 가능 (`--repo kebab --repo other`) OR 매칭. `--code-lang` 는 반복 또는 comma 다중 값 (`--code-lang rust,python`), 알 수 없는 값은 빈 hits. `--media code` 는 Tier 1/2/3 모든 code chunk 포함. 1A-1 시점에서는 indexed 된 code chunk 가 없어 filter 가 항상 빈 결과 — 1A-2 (Rust AST chunker) 머지 이후 실효. |
 | `kebab list docs` | 색인된 문서 목록 |
 | `kebab inspect doc <id>` / `kebab inspect chunk <id>` | raw record 보기 |
@@ -132,7 +132,7 @@ flowchart TB
 
     subgraph Pipeline["도메인 + 파이프라인"]
         parse["parse-md / parse-pdf / parse-image / parse-code"]
-        chunker["chunker (md-heading-v1, pdf-page-v1, code-rust-ast-v1)"]
+        chunker["chunker (md-heading-v1, pdf-page-v1, code-rust-ast-v1, code-python-ast-v1, code-ts-ast-v1, code-js-ast-v1, code-go-ast-v1)"]
         embedder["embedder (fastembed multilingual-e5-large)"]
         retriever["retriever (lexical / vector / hybrid RRF)"]
         rag["RAG pipeline"]

crates/kebab-app/src/fetch.rs

@@ -189,10 +189,12 @@ fn fetch_span(
     // (markdown / note / paper / reference / inbox) is the *user-facing*
     // category, not the rendering format — the actual byte-level format
     // lives on the source `RawAsset.media_type`. Look it up via
-    // workspace_path (unique key per asset).
-    if let Some(asset) = <kebab_store_sqlite::SqliteStore as DocumentStore>::get_asset_by_workspace_path(
+    // doc.source_asset_id (PRIMARY KEY) so twin files (identical content
+    // at different paths) always read *this* document's own asset row,
+    // not whichever twin last wrote `assets.workspace_path`.
+    if let Some(asset) = <kebab_store_sqlite::SqliteStore as DocumentStore>::get_asset(
         &app.sqlite,
-        &doc.workspace_path,
+        &doc.source_asset_id,
     )? {
         if matches!(
             asset.media_type,

crates/kebab-app/src/lib.rs

@@ -39,7 +39,7 @@ use std::sync::Arc;
 use anyhow::{Context, anyhow};
 use serde::{Deserialize, Serialize};
 
-use kebab_chunk::{CodeRustAstV1Chunker, MdHeadingV1Chunker, PdfPageV1Chunker};
+use kebab_chunk::{CodeGoAstV1Chunker, CodeJavaAstV1Chunker, CodeJsAstV1Chunker, CodeKotlinAstV1Chunker, CodePythonAstV1Chunker, CodeRustAstV1Chunker, CodeTsAstV1Chunker, MdHeadingV1Chunker, PdfPageV1Chunker};
 use kebab_core::{
     Answer, Block, CanonicalDocument, Chunk, ChunkId, ChunkPolicy, ChunkerVersion, Chunker,
     DocFilter, DocSummary, DocumentId, DocumentStore, Embedder, EmbeddingInput,
@@ -50,7 +50,7 @@ use kebab_core::{
 use kebab_llm_local::OllamaLanguageModel;
 use kebab_normalize::build_canonical_document;
 use kebab_parse_image::{ImageExtractor, OllamaVisionOcr, apply_caption, apply_ocr};
-use kebab_parse_code::RustAstExtractor;
+use kebab_parse_code::{GoAstExtractor, JavaAstExtractor, JavascriptAstExtractor, KotlinAstExtractor, PythonAstExtractor, RustAstExtractor, TypescriptAstExtractor};
 use kebab_parse_pdf::PdfTextExtractor;
 use kebab_parse_md::{BodyHints, parse_blocks, parse_frontmatter};
 use kebab_source_fs::FsSourceConnector;
@@ -71,7 +71,7 @@ mod staleness;
 
 pub use app::{App, SearchResponse};
 pub use ingest_progress::{AggregateCounts, IngestEvent, render_skipped_breakdown};
-pub use reset::{ResetReport, ResetScope};
+pub use reset::{ResetReport, ResetScope, enumerate_orphans};
 pub use error_wire::{ERROR_V1_ID, ErrorV1, StructuredError, classify};
 pub use fetch::fetch_with_config;
 #[doc(hidden)]
@@ -375,6 +375,28 @@ pub fn ingest_with_config_opts(
         .map(|d| d.doc_id.0)
         .collect();
 
+    // Dogfood: post-walker sweep to remove stored docs whose source
+    // file has been deleted from the filesystem. Must run BEFORE the
+    // per-asset loop so the loop's New/Updated labelling is based on
+    // the post-purge store state (the purged doc_ids won't be in
+    // `existing_doc_ids` above — they were already removed, OR the
+    // sweep here removes them before we start counting).
+    //
+    // Critical design invariant: only purge when the file is TRULY
+    // absent from disk. A file that is still on disk but outside the
+    // current walker scope (config narrowing / include-glob change) is
+    // NOT purged — we leave it in place to protect against accidental
+    // data loss via config edits.
+    let scanned_paths: std::collections::HashSet<kebab_core::WorkspacePath> = assets
+        .iter()
+        .map(|a| a.workspace_path.clone())
+        .collect();
+    let purged_deleted_files = sweep_deleted_files(
+        &app,
+        &scanned_paths,
+        vector_store.as_ref().map(|v| v.as_ref()),
+    )?;
+
     let started_at = time::OffsetDateTime::now_utc();
 
     let mut items: Vec<kebab_core::IngestItem> = Vec::new();
@@ -647,11 +669,11 @@ pub fn ingest_with_config_opts(
     crate::ingest_progress::emit(progress, terminal_event);
 
     // p9-fb-19: bump the persistent corpus_revision counter when a
-    // commit landed (any new / updated). This invalidates every
+    // commit landed (any new / updated / purged). This invalidates every
     // entry in any in-process LRU search cache (in this process or
     // a sibling) on the next lookup. No-op when nothing changed
     // (skipped-only run) — the cache stays valid.
-    if new_count > 0 || updated_count > 0 {
+    if new_count > 0 || updated_count > 0 || purged_deleted_files > 0 {
         match app.sqlite.bump_corpus_revision() {
             Ok(rev) => tracing::debug!(
                 target: "kebab-app",
@@ -682,6 +704,7 @@ pub fn ingest_with_config_opts(
         skipped_generated: fs_skips.skipped_generated,
         skipped_size_exceeded: fs_skips.skipped_size_exceeded,
         skip_examples: fs_skips.skip_examples,
+        purged_deleted_files,
         items: if summary_only { None } else { Some(items) },
     })
 }
@@ -748,15 +771,18 @@ struct ImagePipeline<'a> {
 /// hold (per design §9 cascade rule):
 ///
 /// 1. `force_reingest == false` — caller hasn't asked to bypass skip.
-/// 2. The freshly-scanned asset's blake3 checksum equals what the
-///    existing `assets` row stores at the same `workspace_path`.
-/// 3. The doc keyed on `(workspace_path, asset_id, current_parser_version)`
-///    exists. If the parser_version changed, `id_for_doc` produces a
-///    different `doc_id` so the lookup misses → no skip → re-process.
-/// 4. The existing doc's stamped `last_chunker_version` AND
-///    `last_embedding_version` match the values the caller is about
-///    to use (`Some(v) == Some(v)` and `None == None` — see design
-///    doc for the `None == None` rule when no embedder is configured).
+/// 2. A document already exists at this `workspace_path`
+///    (`get_document_by_workspace_path`). The lookup is document-side, not
+///    asset-side, so twin files (identical content at different paths) each
+///    hit their own stable doc row — `documents.workspace_path` is UNIQUE
+///    while `assets` may dedupe content into a single row with a flip-flop
+///    `workspace_path` column (dogfood bug #4, see `tasks/HOTFIXES.md`).
+/// 3. The existing doc's `source_asset_id` equals the freshly-scanned
+///    asset's blake3 checksum (content unchanged).
+/// 4. The existing doc's `parser_version` matches the current extractor's
+///    `parser_version` (extractor not upgraded). Combined with `chunker_version`
+///    and `last_embedding_version` checks immediately below — full cascade
+///    per design §9.
 ///
 /// Returns `Ok(None)` (proceed with full re-process) when any check
 /// fails or any DB read errors out — the skip path is opportunistic;
@@ -773,31 +799,19 @@ fn try_skip_unchanged(
     if force_reingest {
         return Ok(None);
     }
-    let existing_asset = match app
+    // Document-centric skip: look up the existing document row by
+    // workspace_path directly. This avoids the twin-file flip-flop
+    // that the old asset-side lookup suffers from — multiple files
+    // with identical content share one `assets` row whose
+    // `workspace_path` is overwritten on every UPSERT, so
+    // `get_asset_by_workspace_path(path1)` could return the OTHER
+    // twin's path (or None) after any ingest of the twin. The
+    // `documents` table has a UNIQUE index on `workspace_path` (V001),
+    // so each twin has its own stable row regardless of asset de-dup.
+    let existing_doc = match app
         .sqlite
-        .get_asset_by_workspace_path(&asset.workspace_path)
+        .get_document_by_workspace_path(&asset.workspace_path)
     {
-        Ok(Some(a)) => a,
-        Ok(None) => return Ok(None),
-        Err(e) => {
-            tracing::debug!(
-                target: "kebab-app",
-                path = %asset.workspace_path.0,
-                error = %e,
-                "skip-check: get_asset_by_workspace_path failed; falling through to re-process"
-            );
-            return Ok(None);
-        }
-    };
-    if existing_asset.checksum != asset.checksum {
-        return Ok(None);
-    }
-    let candidate_doc_id = kebab_core::id_for_doc(
-        &asset.workspace_path,
-        &asset.asset_id,
-        current_parser_version,
-    );
-    let existing_doc = match app.sqlite.get_document(&candidate_doc_id) {
         Ok(Some(d)) => d,
         Ok(None) => return Ok(None),
         Err(e) => {
@@ -805,21 +819,37 @@ fn try_skip_unchanged(
                 target: "kebab-app",
                 path = %asset.workspace_path.0,
                 error = %e,
-                "skip-check: get_document failed; falling through to re-process"
+                "skip-check: get_document_by_workspace_path failed; falling through to re-process"
             );
             return Ok(None);
         }
     };
+    // 1. Content unchanged: the freshly-computed asset_id (blake3
+    //    content hash) must match what this document was ingested from.
+    if existing_doc.source_asset_id != asset.asset_id {
+        return Ok(None);
+    }
+    // 2. Parser unchanged: parser_version is baked into id_for_doc so
+    //    a version bump yields a different doc_id and the row above
+    //    would have been missing. Checking here explicitly keeps the
+    //    logic self-documenting and guards against future id_for_doc
+    //    changes.
+    if existing_doc.parser_version != *current_parser_version {
+        return Ok(None);
+    }
+    // 3. Chunker unchanged.
     let chunker_match = existing_doc.last_chunker_version.as_ref()
         == Some(current_chunker_version);
     if !chunker_match {
         return Ok(None);
     }
+    // 4. Embedder unchanged.
     let embedder_match = existing_doc.last_embedding_version.as_ref()
         == current_embedding_version;
     if !embedder_match {
         return Ok(None);
     }
+    let candidate_doc_id = existing_doc.doc_id.clone();
     tracing::debug!(
         target: "kebab-app::ingest",
         path = %asset.workspace_path.0,
@@ -918,8 +948,11 @@ fn ingest_one_asset(
                 force_reingest,
             );
         }
-        // p10-1A-2 Task 8: Rust code ingest.
-        MediaType::Code(lang) if lang == "rust" => {
+        // p10-1A-2 / 1B: code ingest dispatch.
+        MediaType::Code(lang)
+            if matches!(lang.as_str(),
+                "rust" | "python" | "typescript" | "javascript" | "go" | "java" | "kotlin") =>
+        {
             return ingest_one_code_asset(
                 app,
                 asset,
@@ -928,6 +961,7 @@ fn ingest_one_asset(
                 vector_store,
                 existing_doc_ids,
                 force_reingest,
+                lang.as_str(),
             );
         }
         // p10-1A-2: non-Rust Code, Audio, and Other are not yet wired;
@@ -1443,6 +1477,120 @@ fn purge_vector_orphans_for_workspace_path(
     Ok(())
 }
 
+/// Dogfood: post-walker sweep that purges stored documents whose source
+/// file has been physically deleted from the filesystem.
+///
+/// Algorithm:
+/// 1. Query `documents` for every `workspace_path` currently stored.
+/// 2. Compute `orphan_candidates = stored_paths - scanned_paths`.
+/// 3. For each candidate: resolve to an absolute path and call
+///    `Path::try_exists().unwrap_or(true)` — transient FS errors
+///    (EACCES, NFS hiccup, ownership change) conservatively count as
+///    "still present" so we never purge on uncertain signal. If the
+///    file still exists on disk it was merely out-of-scope this run
+///    (config narrowing / include-glob change) — leave it alone. Only
+///    files that are truly absent trigger a purge.
+/// 4. For absent files: call `purge_deleted_workspace_path` (SQLite
+///    cascade delete + optional copied-asset file removal) and, if a
+///    vector store is present, delete the associated vectors.
+///
+/// Returns the number of documents purged.
+///
+/// Non-fatal design: individual purge failures are logged and counted
+/// as errors on the per-file level but do NOT abort the sweep — a
+/// partial failure is preferable to blocking the rest of ingest. The
+/// return value only counts successful purges.
+fn sweep_deleted_files(
+    app: &App,
+    scanned_paths: &std::collections::HashSet<kebab_core::WorkspacePath>,
+    vector_store: Option<&kebab_store_vector::LanceVectorStore>,
+) -> anyhow::Result<u32> {
+    use kebab_core::DocumentStore as _;
+
+    let stored_paths = app
+        .sqlite
+        .all_workspace_paths()
+        .context("sweep_deleted_files: all_workspace_paths")?;
+
+    if stored_paths.is_empty() {
+        return Ok(0);
+    }
+
+    let workspace_root = app.config.resolve_workspace_root();
+    let mut purged: u32 = 0;
+
+    for stored_path in stored_paths {
+        if scanned_paths.contains(&stored_path) {
+            continue; // still in scope — skip
+        }
+
+        // Resolve to an absolute path and check existence on disk.
+        // Use `try_exists` + `unwrap_or(true)` so transient FS errors
+        // (EACCES on a path we lack read on, NFS hiccups, ownership
+        // change) are CONSERVATIVELY treated as "file still present" —
+        // never purge on uncertain signal (data-safety: PR #148 review).
+        // `exists()` would return false on Err and trigger a wrongful
+        // purge. Files whose path cannot be joined (theoretically
+        // impossible for non-empty workspace_path strings, but
+        // defense-in-depth) are likewise treated as still present.
+        let abs = workspace_root.join(&stored_path.0);
+        if abs.try_exists().unwrap_or(true) {
+            // File is on disk but not in this scan's scope (config
+            // narrowing). DO NOT purge — critical design constraint.
+            tracing::debug!(
+                target: "kebab-app",
+                path = %stored_path.0,
+                "sweep_deleted_files: file on disk but out of scope — leaving in store"
+            );
+            continue;
+        }
+
+        // File is truly absent → purge.
+        let chunk_ids = match kebab_store_sqlite::purge_deleted_workspace_path(
+            &app.sqlite,
+            &stored_path,
+        ) {
+            Ok(ids) => ids,
+            Err(e) => {
+                tracing::warn!(
+                    target: "kebab-app",
+                    path = %stored_path.0,
+                    error = %e,
+                    "sweep_deleted_files: purge failed; skipping this path"
+                );
+                continue;
+            }
+        };
+
+        // Purge associated vectors (best-effort; partial failure
+        // acceptable — orphan vectors get cleaned by `kebab reset
+        // --vector-only` if they accumulate).
+        if let Some(vec) = vector_store {
+            if !chunk_ids.is_empty() {
+                use kebab_core::VectorStore as _;
+                if let Err(e) = vec.delete_by_chunk_ids(&chunk_ids) {
+                    tracing::warn!(
+                        target: "kebab-app",
+                        path = %stored_path.0,
+                        count = chunk_ids.len(),
+                        error = %e,
+                        "sweep_deleted_files: vector delete failed; SQLite side already cleaned"
+                    );
+                }
+            }
+        }
+
+        tracing::info!(
+            target: "kebab-app",
+            path = %stored_path.0,
+            "sweep_deleted_files: purged document for deleted file"
+        );
+        purged = purged.saturating_add(1);
+    }
+
+    Ok(purged)
+}
+
 /// P7-3: process one `MediaType::Pdf` asset end-to-end.
 ///
 /// - Reads bytes from disk.
@@ -1642,6 +1790,7 @@ fn ingest_one_pdf_asset(
 ///
 /// All other steps (incremental skip, byte read, ExtractContext, put_*,
 /// embed, purge_vector_orphans) are identical to the PDF function.
+#[allow(clippy::too_many_arguments)]
 fn ingest_one_code_asset(
     app: &App,
     asset: &RawAsset,
@@ -1650,6 +1799,7 @@ fn ingest_one_code_asset(
     vector_store: Option<&Arc<kebab_store_vector::LanceVectorStore>>,
     existing_doc_ids: &std::collections::HashSet<String>,
     force_reingest: bool,
+    code_lang: &str,                        // <-- NEW (p10-1b Task D)
 ) -> anyhow::Result<kebab_core::IngestItem> {
     let path = match &asset.source_uri {
         SourceUri::File(p) => p.clone(),
@@ -1671,17 +1821,36 @@ fn ingest_one_code_asset(
             });
         }
     };
-    // p10-1A-2 task 8: incremental-ingest early-skip for the code flow.
-    // Code docs use `code-rust-v1` as the parser_version and
-    // `CodeRustAstV1Chunker` as the chunker — both pinned per-medium
-    // today (no config knob).
-    let code_parser_version =
-        ParserVersion(kebab_parse_code::RUST_PARSER_VERSION.to_string());
+
+    // p10-1b Task D/G/J: parser_version per-lang.
+    let parser_version = match code_lang {
+        "rust"       => ParserVersion(kebab_parse_code::RUST_PARSER_VERSION.to_string()),
+        "python"     => ParserVersion(kebab_parse_code::PYTHON_PARSER_VERSION.to_string()),
+        "typescript" => ParserVersion(kebab_parse_code::TS_PARSER_VERSION.to_string()),
+        "javascript" => ParserVersion(kebab_parse_code::JS_PARSER_VERSION.to_string()),
+        "go" => ParserVersion(kebab_parse_code::GO_PARSER_VERSION.to_string()),
+        "java" => ParserVersion(kebab_parse_code::JAVA_PARSER_VERSION.to_string()),
+        "kotlin" => ParserVersion(kebab_parse_code::KOTLIN_PARSER_VERSION.to_string()),
+        other => anyhow::bail!("unsupported code_lang: {other}"),
+    };
+
+    // p10-1b Task D/G/J/L: chunker_version per-lang.
+    let chunker_version = match code_lang {
+        "rust"       => CodeRustAstV1Chunker.chunker_version(),
+        "python"     => CodePythonAstV1Chunker.chunker_version(),
+        "typescript" => CodeTsAstV1Chunker.chunker_version(),
+        "javascript" => CodeJsAstV1Chunker.chunker_version(),
+        "go" => CodeGoAstV1Chunker.chunker_version(),
+        "java" => CodeJavaAstV1Chunker.chunker_version(),
+        "kotlin" => CodeKotlinAstV1Chunker.chunker_version(),
+        other => anyhow::bail!("unreachable chunker_version: {other}"),
+    };
+
     if let Some(item) = try_skip_unchanged(
         app,
         asset,
-        &code_parser_version,
-        &CodeRustAstV1Chunker.chunker_version(),
+        &parser_version,
+        &chunker_version,
         embedder.map(|e| e.model_version()).as_ref(),
         force_reingest,
     )? {
@@ -1697,20 +1866,62 @@ fn ingest_one_code_asset(
         workspace_root: &workspace_root,
         config: &extract_config,
     };
-    let mut canonical = RustAstExtractor::new()
-        .extract(&ctx, &bytes)
-        .context("kb-parse-code::RustAstExtractor::extract")?;
 
-    // Per-medium chunker selection: Rust code always uses code-rust-ast-v1
-    // regardless of `config.chunking.chunker_version`.
-    let chunker = CodeRustAstV1Chunker;
-    let chunks = chunker
-        .chunk(&canonical, chunk_policy)
-        .context("kb-chunk::CodeRustAstV1Chunker::chunk")?;
+    // p10-1b Task D/G/J/L: extractor per-lang.
+    let mut canonical = match code_lang {
+        "rust" => RustAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::RustAstExtractor::extract (code:rust)")?,
+        "python" => PythonAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::PythonAstExtractor::extract (code:python)")?,
+        "typescript" => TypescriptAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::TypescriptAstExtractor::extract (code:typescript)")?,
+        "javascript" => JavascriptAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::JavascriptAstExtractor::extract (code:javascript)")?,
+        "go" => GoAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::GoAstExtractor::extract (code:go)")?,
+        "java" => JavaAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::JavaAstExtractor::extract (code:java)")?,
+        "kotlin" => KotlinAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .context("kb-parse-code::KotlinAstExtractor::extract (code:kotlin)")?,
+        other => anyhow::bail!("unreachable (extract): {other}"),
+    };
+
+    // p10-1b Task D/G/J/L: chunker per-lang.
+    let chunks = match code_lang {
+        "rust" => CodeRustAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodeRustAstV1Chunker::chunk (code:rust)")?,
+        "python" => CodePythonAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodePythonAstV1Chunker::chunk (code:python)")?,
+        "typescript" => CodeTsAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodeTsAstV1Chunker::chunk (code:typescript)")?,
+        "javascript" => CodeJsAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodeJsAstV1Chunker::chunk (code:javascript)")?,
+        "go" => CodeGoAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodeGoAstV1Chunker::chunk (code:go)")?,
+        "java" => CodeJavaAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodeJavaAstV1Chunker::chunk (code:java)")?,
+        "kotlin" => CodeKotlinAstV1Chunker
+            .chunk(&canonical, chunk_policy)
+            .context("kb-chunk::CodeKotlinAstV1Chunker::chunk (code:kotlin)")?,
+        other => anyhow::bail!("unreachable (chunk): {other}"),
+    };
 
     // Stamp chunker + embedding versions so incremental skip detection has
     // data on the second run.
-    canonical.last_chunker_version = Some(chunker.chunker_version());
+    canonical.last_chunker_version = Some(chunker_version.clone());
     if let Some(emb) = embedder {
         canonical.last_embedding_version = Some(emb.model_version());
     }
@@ -1794,7 +2005,7 @@ fn ingest_one_code_asset(
         block_count: u32::try_from(canonical.blocks.len()).ok(),
         chunk_count: u32::try_from(chunks.len()).ok(),
         parser_version: Some(canonical.parser_version.clone()),
-        chunker_version: Some(chunker.chunker_version()),
+        chunker_version: Some(chunker_version),
         warnings,
         error: None,
     })

crates/kebab-app/src/reset.rs

@@ -9,13 +9,19 @@
 //!
 //! `--vector-only` additionally truncates `embedding_records` in SQLite
 //! so the next `kebab ingest` re-embeds cleanly without orphan rows.
+//!
+//! `--orphans-only` purges stored docs that are outside the current walker
+//! scope (config narrowing / removed sub-directory). No filesystem paths are
+//! removed — this is purely a store-level reconciliation.
 
+use std::collections::HashSet;
 use std::path::PathBuf;
 
 use anyhow::{Context, Result};
 use serde::{Deserialize, Serialize};
 
 use kebab_config::{Config, expand_path};
+use kebab_core::WorkspacePath;
 
 /// What the user asked to remove. Mutually exclusive — picked by the CLI
 /// from a clap `ArgGroup`.
@@ -32,6 +38,13 @@ pub enum ResetScope {
     VectorOnly,
     /// Wipe only the config dir.
     ConfigOnly,
+    /// Purge stored docs that are outside the current walker scope (no
+    /// filesystem paths are removed). Filesystem existence is NOT checked —
+    /// anything the current walker would not visit is considered an orphan.
+    /// The explicit complement to the conservative `sweep_deleted_files`
+    /// that runs during ingest (which leaves on-disk-but-out-of-scope docs
+    /// alone for data safety).
+    OrphansOnly,
 }
 
 /// Result of a successful wipe — emitted as `reset_report.v1` by the
@@ -41,6 +54,16 @@ pub struct ResetReport {
     pub scope: ResetScope,
     pub removed_paths: Vec<PathBuf>,
     pub embedding_rows_truncated: u64,
+    /// Number of stored docs purged because they are outside the current
+    /// walker scope. Non-zero only when `scope == OrphansOnly`.
+    /// `#[serde(default)]` preserves back-compat with older callers that
+    /// do not include this field.
+    #[serde(default)]
+    pub orphans_purged: u32,
+    /// Paths of the orphaned docs that were purged. Sorted for deterministic
+    /// output. Non-empty only when `scope == OrphansOnly`.
+    #[serde(default)]
+    pub purged_paths: Vec<WorkspacePath>,
 }
 
 /// Compute the absolute on-disk paths a given scope will wipe, given a
@@ -67,6 +90,10 @@ pub fn enumerate_paths(scope: ResetScope, cfg: &Config) -> Vec<PathBuf> {
             vec![vector_dir]
         }
         ResetScope::ConfigOnly => vec![cfg_dir],
+        // OrphansOnly operates purely at the store level — no filesystem paths
+        // are removed. Return empty so `estimate_size_bytes` stays zero and
+        // the existing confirm UI path for directory wipes is skipped.
+        ResetScope::OrphansOnly => vec![],
     }
 }
 
@@ -96,16 +123,82 @@ pub fn estimate_size_bytes(paths: &[PathBuf]) -> u64 {
     paths.iter().map(|p| walk(p)).sum()
 }
 
+/// Compute the workspace paths stored in SQLite that are NOT visited by
+/// the current walker scope (i.e. they are "orphans" — on disk but
+/// outside the configured include/exclude rules, or from a sub-directory
+/// that has since been removed from the workspace).
+///
+/// Does NOT check filesystem existence — `OrphansOnly` is the explicit
+/// "I know what I'm doing" variant; callers that want the conservative
+/// fs-aware sweep should use `sweep_deleted_files` inside ingest.
+///
+/// Returns the list sorted for deterministic output. Called twice by the
+/// CLI path (once for the confirm UI preview, once inside `execute`);
+/// the double scan is acceptable for a rare destructive operation.
+pub fn enumerate_orphans(cfg: &Config) -> Result<Vec<WorkspacePath>> {
+    use kebab_core::DocumentStore as _;
+    use kebab_source_fs::FsSourceConnector;
+    use kebab_core::SourceScope;
+
+    let store = kebab_store_sqlite::SqliteStore::open(cfg)
+        .context("enumerate_orphans: open SqliteStore")?;
+
+    let stored = store
+        .all_workspace_paths()
+        .context("enumerate_orphans: all_workspace_paths")?;
+
+    if stored.is_empty() {
+        return Ok(Vec::new());
+    }
+
+    // Build the same SourceScope the CLI's ingest path uses: root from
+    // config, exclude list from config, no include override (full scope).
+    let root = cfg.resolve_workspace_root();
+    let scope = SourceScope {
+        root: root.clone(),
+        exclude: cfg.workspace.exclude.clone(),
+        ..Default::default()
+    };
+
+    let connector = FsSourceConnector::new(cfg)
+        .context("enumerate_orphans: build FsSourceConnector")?;
+    let (assets, _skips) = connector
+        .scan_with_skips(&scope)
+        .context("enumerate_orphans: scan workspace")?;
+
+    let scanned: HashSet<WorkspacePath> = assets
+        .into_iter()
+        .map(|a| a.workspace_path)
+        .collect();
+
+    let mut orphans: Vec<WorkspacePath> = stored
+        .into_iter()
+        .filter(|p| !scanned.contains(p))
+        .collect();
+    orphans.sort_by(|a, b| a.0.cmp(&b.0));
+    Ok(orphans)
+}
+
 /// Wipe every path from `enumerate_paths(scope, cfg)`. For
 /// `ResetScope::VectorOnly`, also truncates the SQLite
 /// `embedding_records` table so the store doesn't point at the Lance
 /// rows we just removed off-disk.
 ///
+/// For `ResetScope::OrphansOnly`, no filesystem directories are removed.
+/// Instead the store is reconciled: stored docs outside the current walker
+/// scope are purged from SQLite (+ vector store when configured). The
+/// caller is expected to have already shown the confirm UI using
+/// `enumerate_orphans`.
+///
 /// Idempotent: a missing path is treated as already-removed (success).
 /// Returns a `ResetReport` listing exactly what was removed (paths that
 /// existed before the call) so `--json` callers see the truth, not the
 /// request.
 pub fn execute(scope: ResetScope, cfg: &Config) -> Result<ResetReport> {
+    if matches!(scope, ResetScope::OrphansOnly) {
+        return execute_orphans_only(cfg);
+    }
+
     let paths = enumerate_paths(scope, cfg);
     let mut removed = Vec::new();
 
@@ -128,9 +221,100 @@ pub fn execute(scope: ResetScope, cfg: &Config) -> Result<ResetReport> {
         scope,
         removed_paths: removed,
         embedding_rows_truncated,
+        orphans_purged: 0,
+        purged_paths: Vec::new(),
     })
 }
 
+/// Execute the `OrphansOnly` variant: reconcile stored docs against the
+/// current walker scope without touching any filesystem directory.
+fn execute_orphans_only(cfg: &Config) -> Result<ResetReport> {
+    let orphans = enumerate_orphans(cfg)
+        .context("execute_orphans_only: enumerate orphans")?;
+
+    if orphans.is_empty() {
+        return Ok(ResetReport {
+            scope: ResetScope::OrphansOnly,
+            removed_paths: Vec::new(),
+            embedding_rows_truncated: 0,
+            orphans_purged: 0,
+            purged_paths: Vec::new(),
+        });
+    }
+
+    let store = std::sync::Arc::new(
+        kebab_store_sqlite::SqliteStore::open(cfg)
+            .context("execute_orphans_only: open SqliteStore")?,
+    );
+
+    // Open vector store if configured. Mirror the same guard the ingest
+    // path uses: only construct when the provider is not "none" / dims > 0.
+    let vector_store: Option<kebab_store_vector::LanceVectorStore> =
+        open_vector_store_if_configured(cfg, store.clone())?;
+
+    let mut purged_paths: Vec<WorkspacePath> = Vec::new();
+
+    for path in &orphans {
+        let chunk_ids = kebab_store_sqlite::purge_deleted_workspace_path(&store, path)
+            .with_context(|| format!("execute_orphans_only: purge {}", path.0))?;
+
+        if let Some(ref vs) = vector_store {
+            if !chunk_ids.is_empty() {
+                use kebab_core::VectorStore as _;
+                if let Err(e) = vs.delete_by_chunk_ids(&chunk_ids) {
+                    tracing::warn!(
+                        target: "kebab-app",
+                        path = %path.0,
+                        count = chunk_ids.len(),
+                        error = %e,
+                        "reset --orphans-only: vector delete failed; SQLite side already cleaned"
+                    );
+                }
+            }
+        }
+
+        tracing::info!(
+            target: "kebab-app",
+            path = %path.0,
+            "reset --orphans-only: purged orphan document"
+        );
+        purged_paths.push(path.clone());
+    }
+
+    let orphans_purged = u32::try_from(purged_paths.len()).unwrap_or(u32::MAX);
+
+    Ok(ResetReport {
+        scope: ResetScope::OrphansOnly,
+        removed_paths: Vec::new(),
+        embedding_rows_truncated: 0,
+        orphans_purged,
+        purged_paths,
+    })
+}
+
+/// Open the Lance vector store if the configured embedding provider is
+/// active (non-"none", dimensions > 0). Returns `None` for lexical-only
+/// configs. Mirrors the guard in `App::vector`.
+fn open_vector_store_if_configured(
+    cfg: &Config,
+    store: std::sync::Arc<kebab_store_sqlite::SqliteStore>,
+) -> Result<Option<kebab_store_vector::LanceVectorStore>> {
+    if cfg.models.embedding.provider == "none" || cfg.models.embedding.dimensions == 0 {
+        return Ok(None);
+    }
+    match kebab_store_vector::LanceVectorStore::new(cfg, store) {
+        Ok(vs) => Ok(Some(vs)),
+        Err(e) => {
+            tracing::warn!(
+                target: "kebab-app",
+                error = %e,
+                "reset --orphans-only: could not open vector store; skipping vector delete"
+            );
+            Ok(None)
+        }
+    }
+}
+
 /// Open the SQLite store at the configured path and run
 /// `truncate_embedding_records`. Returns the count of truncated rows
 /// (the helper itself reports `DELETE` rowcount). If the SQLite file
@@ -200,4 +384,14 @@ mod tests {
         let bytes = estimate_size_bytes(&[dir.path().to_path_buf()]);
         assert_eq!(bytes, 5 + 6);
     }
+
+    #[test]
+    fn enumerate_orphans_only_returns_empty_paths() {
+        let cfg = Config::defaults();
+        let paths = enumerate_paths(ResetScope::OrphansOnly, &cfg);
+        assert!(
+            paths.is_empty(),
+            "OrphansOnly must return empty vec from enumerate_paths"
+        );
+    }
 }

crates/kebab-app/src/schema.rs

@@ -168,7 +168,9 @@ fn collect_stats(
         stale_doc_count: counts.stale_doc_count,
         // p10-1A-2: populated by the store query added in this task.
         code_lang_breakdown: store.code_lang_breakdown()?,
-        repo_breakdown: std::collections::BTreeMap::new(),
+        // p10-1A-2 follow-up: dogfooding (2026-05-20) revealed this was a
+        // placeholder — mirror of code_lang_breakdown for the repo field.
+        repo_breakdown: store.repo_breakdown()?,
     })
 }
 

crates/kebab-app/tests/code_ingest_smoke.rs

@@ -159,6 +159,450 @@ fn rust_code_search_hit_has_repo() {
     );
 }
 
+/// p10-1b Task G: a `.py` file in a sub-directory is ingested and the
+/// resulting `Citation::Code` hit must carry `lang="python"`,
+/// `symbol="kebab_eval.metrics.compute_mrr"`, and `line_start >= 1`.
+/// The sub-directory (`kebab_eval/`) ensures `module_path_for_python`
+/// produces a non-empty prefix so the fully-qualified symbol assertion
+/// exercises the prefix wiring end-to-end.
+#[test]
+fn python_file_ingests_and_searches_as_code_citation() {
+    let env = TestEnv::lexical_only();
+
+    let module_dir = env.workspace_root.join("kebab_eval");
+    std::fs::create_dir_all(&module_dir).unwrap();
+    std::fs::write(
+        module_dir.join("metrics.py"),
+        "\"\"\"compute metrics.\"\"\"\ndef compute_mrr(scores):\n    return sum(scores) / max(len(scores), 1)\n",
+    )
+    .unwrap();
+
+    let report =
+        kebab_app::ingest_with_config(env.config.clone(), env.scope(), false)
+            .expect("ingest must succeed");
+
+    assert!(report.new >= 1, "python file ingested: {report:?}");
+
+    let items = report.items.as_ref().expect("items present");
+    let py_item = items
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("metrics.py"))
+        .expect("metrics.py item");
+    assert_eq!(
+        py_item.parser_version.as_ref().map(|p| p.0.as_str()),
+        Some("code-python-v1"),
+        "parser_version must be code-python-v1"
+    );
+    assert_eq!(
+        py_item.chunker_version.as_ref().map(|c| c.0.as_str()),
+        Some("code-python-ast-v1"),
+        "chunker_version must be code-python-ast-v1"
+    );
+
+    let hits = kebab_app::search_with_config(env.config.clone(), lexical_query("compute_mrr"))
+        .expect("search must succeed");
+
+    let h = hits
+        .iter()
+        .find(|h| matches!(&h.citation, Citation::Code { .. }))
+        .expect("at least one Citation::Code hit for 'compute_mrr'");
+
+    match &h.citation {
+        Citation::Code {
+            lang,
+            symbol,
+            line_start,
+            ..
+        } => {
+            assert_eq!(
+                lang.as_deref(),
+                Some("python"),
+                "citation.lang must be 'python'"
+            );
+            assert_eq!(
+                symbol.as_deref(),
+                Some("kebab_eval.metrics.compute_mrr"),
+                "citation.symbol must be 'kebab_eval.metrics.compute_mrr'"
+            );
+            assert!(*line_start >= 1, "line_start must be >=1");
+        }
+        _ => unreachable!(),
+    }
+
+    assert_eq!(
+        h.code_lang.as_deref(),
+        Some("python"),
+        "SearchHit.code_lang must be 'python'"
+    );
+}
+
+/// p10-1b Task J: a `.ts` file in a sub-directory is ingested and the
+/// resulting `Citation::Code` hit must carry `lang="typescript"`,
+/// `symbol="src/Foo.Foo.bar"`, and `line_start >= 1`.
+/// The sub-directory (`src/`) ensures `module_path_for_tsjs` produces
+/// a non-empty prefix so the fully-qualified symbol assertion exercises
+/// the prefix wiring end-to-end.
+#[test]
+fn typescript_file_ingests_and_searches_as_code_citation() {
+    let env = TestEnv::lexical_only();
+
+    let src_dir = env.workspace_root.join("src");
+    std::fs::create_dir_all(&src_dir).unwrap();
+    std::fs::write(
+        src_dir.join("Foo.ts"),
+        "export class Foo {\n    bar(): number { return 42; }\n}\n",
+    )
+    .unwrap();
+
+    let report =
+        kebab_app::ingest_with_config(env.config.clone(), env.scope(), false)
+            .expect("ingest must succeed");
+
+    assert!(report.new >= 1, "ts file ingested: {report:?}");
+
+    let items = report.items.as_ref().expect("items present");
+    let ts_item = items
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("Foo.ts"))
+        .expect("Foo.ts item");
+    assert_eq!(
+        ts_item.parser_version.as_ref().map(|p| p.0.as_str()),
+        Some("code-ts-v1"),
+        "parser_version must be code-ts-v1"
+    );
+    assert_eq!(
+        ts_item.chunker_version.as_ref().map(|c| c.0.as_str()),
+        Some("code-ts-ast-v1"),
+        "chunker_version must be code-ts-ast-v1"
+    );
+
+    let hits = kebab_app::search_with_config(env.config.clone(), lexical_query("bar"))
+        .expect("search must succeed");
+
+    let h = hits
+        .iter()
+        .find(|h| matches!(&h.citation, Citation::Code { .. }))
+        .expect("at least one Citation::Code hit for 'bar'");
+
+    match &h.citation {
+        Citation::Code {
+            lang,
+            symbol,
+            line_start,
+            ..
+        } => {
+            assert_eq!(
+                lang.as_deref(),
+                Some("typescript"),
+                "citation.lang must be 'typescript'"
+            );
+            assert_eq!(
+                symbol.as_deref(),
+                Some("src/Foo.Foo.bar"),
+                "citation.symbol must be 'src/Foo.Foo.bar'"
+            );
+            assert!(*line_start >= 1, "line_start must be >=1");
+        }
+        _ => unreachable!(),
+    }
+
+    assert_eq!(
+        h.code_lang.as_deref(),
+        Some("typescript"),
+        "SearchHit.code_lang must be 'typescript'"
+    );
+}
+
+/// p10-1b Task L: a `.js` file in a sub-directory is ingested and the
+/// resulting `Citation::Code` hit must carry `lang="javascript"`,
+/// `symbol="src/Bar.Bar.baz"`, and `line_start >= 1`.
+/// The sub-directory (`src/`) ensures `module_path_for_tsjs` produces
+/// a non-empty prefix so the fully-qualified symbol assertion exercises
+/// the prefix wiring end-to-end.
+#[test]
+fn javascript_file_ingests_and_searches_as_code_citation() {
+    let env = TestEnv::lexical_only();
+
+    let src_dir = env.workspace_root.join("src");
+    std::fs::create_dir_all(&src_dir).unwrap();
+    std::fs::write(
+        src_dir.join("Bar.js"),
+        "export class Bar {\n    baz() { return 7; }\n}\n",
+    )
+    .unwrap();
+
+    let report =
+        kebab_app::ingest_with_config(env.config.clone(), env.scope(), false)
+            .expect("ingest must succeed");
+
+    assert!(report.new >= 1, "js file ingested: {report:?}");
+
+    let items = report.items.as_ref().expect("items present");
+    let js_item = items
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("Bar.js"))
+        .expect("Bar.js item");
+    assert_eq!(
+        js_item.parser_version.as_ref().map(|p| p.0.as_str()),
+        Some("code-js-v1"),
+        "parser_version must be code-js-v1"
+    );
+    assert_eq!(
+        js_item.chunker_version.as_ref().map(|c| c.0.as_str()),
+        Some("code-js-ast-v1"),
+        "chunker_version must be code-js-ast-v1"
+    );
+
+    let hits = kebab_app::search_with_config(env.config.clone(), lexical_query("baz"))
+        .expect("search must succeed");
+
+    let h = hits
+        .iter()
+        .find(|h| matches!(&h.citation, Citation::Code { .. }))
+        .expect("at least one Citation::Code hit for 'baz'");
+
+    match &h.citation {
+        Citation::Code {
+            lang,
+            symbol,
+            line_start,
+            ..
+        } => {
+            assert_eq!(
+                lang.as_deref(),
+                Some("javascript"),
+                "citation.lang must be 'javascript'"
+            );
+            assert_eq!(
+                symbol.as_deref(),
+                Some("src/Bar.Bar.baz"),
+                "citation.symbol must be 'src/Bar.Bar.baz'"
+            );
+            assert!(*line_start >= 1, "line_start must be >=1");
+        }
+        _ => unreachable!(),
+    }
+
+    assert_eq!(
+        h.code_lang.as_deref(),
+        Some("javascript"),
+        "SearchHit.code_lang must be 'javascript'"
+    );
+}
+
+/// p10-1c-go Task F: a `.go` file in a sub-directory is ingested and the
+/// resulting `Citation::Code` hit must carry `lang="go"`,
+/// `symbol="chunk.ParseDoc"`, and `line_start >= 1`.
+/// The sub-directory (`chunk/`) ensures the Go package-prefix wiring
+/// produces a non-empty module prefix so the fully-qualified symbol assertion
+/// exercises that path end-to-end.
+#[test]
+fn go_file_ingests_and_searches_as_code_citation() {
+    let env = TestEnv::lexical_only();
+
+    let pkg_dir = env.workspace_root.join("chunk");
+    std::fs::create_dir_all(&pkg_dir).unwrap();
+    std::fs::write(
+        pkg_dir.join("ast.go"),
+        "package chunk\n\nfunc ParseDoc(input string) string {\n    return input\n}\n",
+    )
+    .unwrap();
+
+    let report = kebab_app::ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("ingest must succeed");
+    assert_eq!(report.errors, 0);
+    assert!(report.new >= 1);
+
+    let go_item = report
+        .items
+        .as_ref()
+        .expect("items present")
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("ast.go"))
+        .expect("ast.go item present");
+    assert_eq!(
+        go_item.parser_version.as_ref().map(|p| p.0.as_str()),
+        Some("code-go-v1"),
+        "parser_version must be code-go-v1"
+    );
+    assert_eq!(
+        go_item.chunker_version.as_ref().map(|c| c.0.as_str()),
+        Some("code-go-ast-v1"),
+        "chunker_version must be code-go-ast-v1"
+    );
+
+    let hits = kebab_app::search_with_config(env.config.clone(), lexical_query("ParseDoc"))
+        .expect("search must succeed");
+    let h = hits
+        .iter()
+        .find(|h| matches!(&h.citation, kebab_core::Citation::Code { .. }))
+        .expect("Citation::Code hit");
+    match &h.citation {
+        kebab_core::Citation::Code {
+            lang,
+            symbol,
+            line_start,
+            ..
+        } => {
+            assert_eq!(lang.as_deref(), Some("go"), "citation.lang must be 'go'");
+            assert_eq!(
+                symbol.as_deref(),
+                Some("chunk.ParseDoc"),
+                "citation.symbol must be 'chunk.ParseDoc'"
+            );
+            assert!(*line_start >= 1, "line_start must be >=1");
+        }
+        _ => unreachable!(),
+    }
+    assert_eq!(
+        h.code_lang.as_deref(),
+        Some("go"),
+        "SearchHit.code_lang must be 'go'"
+    );
+}
+
+/// p10-1c-jk Task F: a `.java` file in a package directory is ingested and the
+/// resulting `Citation::Code` hit must carry `lang="java"`,
+/// `symbol="com.foo.Foo.bar"`, and `line_start >= 1`.
+/// The sub-directory (`com/foo/`) ensures the Java package-prefix wiring
+/// produces a non-empty module prefix so the fully-qualified symbol assertion
+/// exercises that path end-to-end.
+#[test]
+fn java_file_ingests_and_searches_as_code_citation() {
+    let env = TestEnv::lexical_only();
+
+    let pkg_dir = env.workspace_root.join("com").join("foo");
+    std::fs::create_dir_all(&pkg_dir).unwrap();
+    std::fs::write(
+        pkg_dir.join("Foo.java"),
+        "package com.foo;\n\npublic class Foo {\n    public String bar() { return \"x\"; }\n}\n",
+    )
+    .unwrap();
+
+    let report = kebab_app::ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("ingest must succeed");
+    assert_eq!(report.errors, 0);
+    assert!(report.new >= 1);
+
+    let java_item = report
+        .items
+        .as_ref()
+        .expect("items present")
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("Foo.java"))
+        .expect("Foo.java item present");
+    assert_eq!(
+        java_item.parser_version.as_ref().map(|p| p.0.as_str()),
+        Some("code-java-v1"),
+        "parser_version must be code-java-v1"
+    );
+    assert_eq!(
+        java_item.chunker_version.as_ref().map(|c| c.0.as_str()),
+        Some("code-java-ast-v1"),
+        "chunker_version must be code-java-ast-v1"
+    );
+
+    let hits = kebab_app::search_with_config(env.config.clone(), lexical_query("bar"))
+        .expect("search must succeed");
+    let h = hits
+        .iter()
+        .find(|h| matches!(&h.citation, kebab_core::Citation::Code { .. }))
+        .expect("Citation::Code hit");
+    match &h.citation {
+        kebab_core::Citation::Code {
+            lang,
+            symbol,
+            line_start,
+            ..
+        } => {
+            assert_eq!(lang.as_deref(), Some("java"), "citation.lang must be 'java'");
+            assert_eq!(
+                symbol.as_deref(),
+                Some("com.foo.Foo.bar"),
+                "citation.symbol must be 'com.foo.Foo.bar'"
+            );
+            assert!(*line_start >= 1, "line_start must be >=1");
+        }
+        _ => unreachable!(),
+    }
+    assert_eq!(
+        h.code_lang.as_deref(),
+        Some("java"),
+        "SearchHit.code_lang must be 'java'"
+    );
+}
+
+/// p10-1c-jk Task I: a `.kt` file in a package directory is ingested and the
+/// resulting `Citation::Code` hit must carry `lang="kotlin"`,
+/// `symbol="com.foo.Foo.bar"`, and `line_start >= 1`.
+/// The sub-directory (`com/foo/`) ensures the Kotlin package-prefix wiring
+/// produces a non-empty module prefix so the fully-qualified symbol assertion
+/// exercises that path end-to-end.
+#[test]
+fn kotlin_file_ingests_and_searches_as_code_citation() {
+    let env = TestEnv::lexical_only();
+
+    let pkg_dir = env.workspace_root.join("com").join("foo");
+    std::fs::create_dir_all(&pkg_dir).unwrap();
+    std::fs::write(
+        pkg_dir.join("Foo.kt"),
+        "package com.foo\n\nclass Foo {\n    fun bar(): String = \"x\"\n}\n",
+    )
+    .unwrap();
+
+    let report = kebab_app::ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("ingest must succeed");
+    assert_eq!(report.errors, 0);
+    assert!(report.new >= 1);
+
+    let kt_item = report
+        .items
+        .as_ref()
+        .expect("items present")
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("Foo.kt"))
+        .expect("Foo.kt item present");
+    assert_eq!(
+        kt_item.parser_version.as_ref().map(|p| p.0.as_str()),
+        Some("code-kotlin-v1"),
+        "parser_version must be code-kotlin-v1"
+    );
+    assert_eq!(
+        kt_item.chunker_version.as_ref().map(|c| c.0.as_str()),
+        Some("code-kotlin-ast-v1"),
+        "chunker_version must be code-kotlin-ast-v1"
+    );
+
+    let hits = kebab_app::search_with_config(env.config.clone(), lexical_query("bar"))
+        .expect("search must succeed");
+    let h = hits
+        .iter()
+        .find(|h| matches!(&h.citation, kebab_core::Citation::Code { .. }))
+        .expect("Citation::Code hit");
+    match &h.citation {
+        kebab_core::Citation::Code {
+            lang,
+            symbol,
+            line_start,
+            ..
+        } => {
+            assert_eq!(lang.as_deref(), Some("kotlin"), "citation.lang must be 'kotlin'");
+            assert_eq!(
+                symbol.as_deref(),
+                Some("com.foo.Foo.bar"),
+                "citation.symbol must be 'com.foo.Foo.bar'"
+            );
+            assert!(*line_start >= 1, "line_start must be >=1");
+        }
+        _ => unreachable!(),
+    }
+    assert_eq!(
+        h.code_lang.as_deref(),
+        Some("kotlin"),
+        "SearchHit.code_lang must be 'kotlin'"
+    );
+}
+
 /// Re-ingesting the same `.rs` file without changes must report
 /// `Unchanged` (incremental-skip path exercised).
 #[test]

crates/kebab-app/tests/file_deletion_auto_purge.rs

@@ -0,0 +1,178 @@
+//! Dogfood: auto-purge stored docs for filesystem-deleted files.
+//!
+//! Two tests:
+//!
+//! 1. `file_deletion_auto_purge` — ingest 2 files, delete one, re-ingest.
+//!    The re-ingest must report `purged_deleted_files = 1`, the deleted
+//!    file must no longer appear in `list_docs`, and lexical search for
+//!    its unique content must return no hits.
+//!
+//! 2. `include_scope_narrowing_does_not_purge` — ingest 2 files under a
+//!    wide glob, narrow the walker scope to only one file, re-ingest.
+//!    The narrowed ingest must NOT purge the out-of-scope file because
+//!    the file is still on disk (just excluded from this run). Protects
+//!    users against accidental data loss via config edits.
+
+mod common;
+
+use common::TestEnv;
+use kebab_app::ingest_with_config_opts;
+use kebab_app::IngestOpts;
+use kebab_core::{DocFilter, DocumentStore, SearchMode, SearchQuery, SourceScope};
+
+/// Helper: open the store via `TestEnv` and run `list_documents`.
+fn list_doc_paths(env: &TestEnv) -> Vec<String> {
+    use kebab_store_sqlite::SqliteStore;
+    let store = SqliteStore::open(&env.config).unwrap();
+    store.run_migrations().unwrap();
+    store
+        .list_documents(&DocFilter::default())
+        .unwrap()
+        .into_iter()
+        .map(|d| d.doc_path.0)
+        .collect()
+}
+
+#[test]
+fn file_deletion_auto_purge() {
+    let env = TestEnv::lexical_only();
+
+    // Write two .rs files into the workspace.
+    let a_path = env.workspace_root.join("a.rs");
+    let b_path = env.workspace_root.join("b.rs");
+    std::fs::write(&a_path, "// file a\nfn alpha() {}\n").unwrap();
+    std::fs::write(&b_path, "// file b\nfn bravo() {}\n").unwrap();
+
+    // First ingest — both must be New.
+    let first = ingest_with_config_opts(
+        env.config.clone(),
+        env.scope(),
+        false,
+        IngestOpts::default(),
+    )
+    .expect("first ingest must succeed");
+    // Only count the .rs files we added (there may be fixture files too).
+    let first_new = first.new;
+    assert!(first_new >= 2, "expected at least 2 new docs: {first:?}");
+    assert_eq!(
+        first.purged_deleted_files, 0,
+        "no purges on first ingest: {first:?}"
+    );
+    assert_eq!(first.errors, 0, "no errors on first ingest: {first:?}");
+
+    // Delete one file from the filesystem.
+    std::fs::remove_file(&b_path).expect("remove b.rs");
+
+    // Second ingest — scanned count drops by 1; b.rs should be purged.
+    let second = ingest_with_config_opts(
+        env.config.clone(),
+        env.scope(),
+        false,
+        IngestOpts::default(),
+    )
+    .expect("second ingest must succeed");
+
+    assert_eq!(
+        second.purged_deleted_files, 1,
+        "exactly 1 file should be purged: {second:?}"
+    );
+    assert_eq!(second.new, 0, "no new docs after deletion: {second:?}");
+    assert_eq!(second.updated, 0, "no updated docs: {second:?}");
+    assert_eq!(second.errors, 0, "no errors: {second:?}");
+
+    // b.rs must no longer appear in list_docs.
+    let doc_paths = list_doc_paths(&env);
+    let b_ws_path = "b.rs";
+    assert!(
+        !doc_paths.iter().any(|p| p == b_ws_path),
+        "b.rs must be gone from list_docs; got: {doc_paths:?}"
+    );
+    // a.rs must still be present.
+    let a_ws_path = "a.rs";
+    assert!(
+        doc_paths.iter().any(|p| p == a_ws_path),
+        "a.rs must still be in list_docs; got: {doc_paths:?}"
+    );
+
+    // Lexical search for b.rs's unique content returns no hits.
+    let app = env.app();
+    let query = SearchQuery {
+        text: "bravo".to_string(),
+        mode: SearchMode::Lexical,
+        k: 10,
+        filters: kebab_core::SearchFilters::default(),
+    };
+    let hits = app.search(query).expect("search must not error");
+    assert!(
+        hits.is_empty(),
+        "search for deleted file's content must return no hits; got: {hits:?}"
+    );
+}
+
+#[test]
+fn include_scope_narrowing_does_not_purge() {
+    let env = TestEnv::lexical_only();
+
+    // Write two .rs files.
+    let a_path = env.workspace_root.join("a_narrow.rs");
+    let b_path = env.workspace_root.join("b_narrow.rs");
+    std::fs::write(&a_path, "// narrow a\nfn alpha_narrow() {}\n").unwrap();
+    std::fs::write(&b_path, "// narrow b\nfn bravo_narrow() {}\n").unwrap();
+
+    // Wide scope: first ingest — both must be New.
+    let wide_scope = SourceScope {
+        root: env.workspace_root.clone(),
+        include: vec!["**/*.rs".to_string()],
+        exclude: env.config.workspace.exclude.clone(),
+    };
+    let first = ingest_with_config_opts(
+        env.config.clone(),
+        wide_scope,
+        false,
+        IngestOpts::default(),
+    )
+    .expect("first ingest (wide) must succeed");
+    assert!(
+        first.new >= 2,
+        "expected at least 2 new docs: {first:?}"
+    );
+    assert_eq!(
+        first.purged_deleted_files, 0,
+        "no purges on first ingest: {first:?}"
+    );
+
+    // Narrow scope: only a_narrow.rs in include — b_narrow.rs is still
+    // on disk but excluded from the walker scope.
+    let narrow_scope = SourceScope {
+        root: env.workspace_root.clone(),
+        include: vec!["a_narrow.rs".to_string()],
+        exclude: env.config.workspace.exclude.clone(),
+    };
+    let second = ingest_with_config_opts(
+        env.config.clone(),
+        narrow_scope,
+        false,
+        IngestOpts::default(),
+    )
+    .expect("second ingest (narrow) must succeed");
+
+    // CRITICAL: b_narrow.rs is still on disk — must NOT be purged.
+    assert_eq!(
+        second.purged_deleted_files, 0,
+        "scope-narrowing must NOT purge on-disk files; got: {second:?}"
+    );
+    assert_eq!(second.errors, 0, "no errors: {second:?}");
+
+    // b_narrow.rs must still exist in the store.
+    let doc_paths = list_doc_paths(&env);
+    let b_ws_path = "b_narrow.rs";
+    assert!(
+        doc_paths.iter().any(|p| p == b_ws_path),
+        "b_narrow.rs must still be in list_docs after scope narrowing; got: {doc_paths:?}"
+    );
+    // And the file must still be on disk.
+    assert!(
+        b_path.exists(),
+        "b_narrow.rs must still be on disk (we didn't delete it)"
+    );
+}

crates/kebab-app/tests/reset_orphans.rs

@@ -0,0 +1,141 @@
+//! Integration test for `kebab reset --orphans-only`.
+//!
+//! Verifies that stored docs outside the current walker scope are purged
+//! from the store without removing any files from the filesystem.
+//!
+//! Test outline:
+//! 1. Ingest 3 .rs files (a.rs, b.rs, c.rs) — all New.
+//! 2. Narrow the config `include` to `["a.rs"]` only; b.rs and c.rs are
+//!    still on disk but outside the walker scope.
+//! 3. Run `execute(ResetScope::OrphansOnly, &cfg)` — report must show
+//!    `orphans_purged == 2` and `purged_paths` contains b.rs + c.rs.
+//! 4. `list docs` must show only a.rs.
+//! 5. b.rs and c.rs must still exist on disk (no filesystem removal).
+//! 6. Second reset → `orphans_purged == 0` (idempotent).
+
+mod common;
+
+use common::TestEnv;
+use kebab_app::IngestOpts;
+use kebab_app::reset::{ResetScope, execute};
+use kebab_core::{DocFilter, DocumentStore, SourceScope};
+
+/// Open the SqliteStore and list all `workspace_path` values.
+fn list_doc_paths(env: &TestEnv) -> Vec<String> {
+    use kebab_store_sqlite::SqliteStore;
+    let store = SqliteStore::open(&env.config).unwrap();
+    store.run_migrations().unwrap();
+    store
+        .list_documents(&DocFilter::default())
+        .unwrap()
+        .into_iter()
+        .map(|d| d.doc_path.0)
+        .collect()
+}
+
+#[test]
+fn reset_orphans_only_purges_out_of_scope_docs() {
+    let env = TestEnv::lexical_only();
+
+    // Write three .rs files into the workspace.
+    let a_path = env.workspace_root.join("a.rs");
+    let b_path = env.workspace_root.join("b.rs");
+    let c_path = env.workspace_root.join("c.rs");
+    std::fs::write(&a_path, "// file a\nfn alpha() {}\n").unwrap();
+    std::fs::write(&b_path, "// file b\nfn bravo() {}\n").unwrap();
+    std::fs::write(&c_path, "// file c\nfn charlie() {}\n").unwrap();
+
+    // Ingest all three with a wide scope.
+    let wide_scope = SourceScope {
+        root: env.workspace_root.clone(),
+        include: vec!["**/*.rs".to_string()],
+        exclude: env.config.workspace.exclude.clone(),
+    };
+    let first = kebab_app::ingest_with_config_opts(
+        env.config.clone(),
+        wide_scope,
+        false,
+        IngestOpts::default(),
+    )
+    .expect("first ingest must succeed");
+    // The fixture workspace may contain other .rs files — just assert we
+    // got at least 3 new docs (our a.rs, b.rs, c.rs).
+    assert!(first.new >= 3, "expected at least 3 new docs: {first:?}");
+    assert_eq!(first.errors, 0, "no errors on first ingest");
+
+    // Narrow config to include only a.rs; b.rs + c.rs are still on disk.
+    let mut narrow_cfg = env.config.clone();
+    narrow_cfg.workspace.exclude.clear();
+    // Re-point workspace root (already correct) and restrict include via
+    // the SourceScope in the connector. The config's `workspace.root` is
+    // used by `enumerate_orphans` to build its scope — we keep that
+    // pointing at the workspace root. We simulate narrowing by setting a
+    // glob that only matches a.rs.
+    //
+    // NOTE: `kebab_config::WorkspaceCfg` does not have an `include` field
+    // (it was removed in p9-fb-25). We narrow the scope via the walker
+    // exclude list: exclude b.rs and c.rs explicitly.
+    narrow_cfg.workspace.exclude = vec!["b.rs".to_string(), "c.rs".to_string()];
+
+    // Run orphans-only reset.
+    let report = execute(ResetScope::OrphansOnly, &narrow_cfg)
+        .expect("orphans-only reset must succeed");
+
+    assert_eq!(
+        report.orphans_purged, 2,
+        "expected 2 orphans purged (b.rs + c.rs): {report:?}"
+    );
+
+    let mut purged: Vec<String> = report
+        .purged_paths
+        .iter()
+        .map(|p| p.0.clone())
+        .collect();
+    purged.sort();
+    assert_eq!(
+        purged,
+        vec!["b.rs".to_string(), "c.rs".to_string()],
+        "purged_paths must list b.rs and c.rs in sorted order: {purged:?}"
+    );
+
+    // list docs must show only a.rs (and any pre-existing fixture files
+    // that are not excluded by the narrow config).
+    let doc_paths = list_doc_paths(&env);
+    // The narrow_cfg excludes b.rs + c.rs — they must no longer be in store.
+    assert!(
+        !doc_paths.iter().any(|p| p == "b.rs"),
+        "b.rs must be gone from store after orphans-only reset; got: {doc_paths:?}"
+    );
+    assert!(
+        !doc_paths.iter().any(|p| p == "c.rs"),
+        "c.rs must be gone from store after orphans-only reset; got: {doc_paths:?}"
+    );
+    assert!(
+        doc_paths.iter().any(|p| p == "a.rs"),
+        "a.rs must still be in store; got: {doc_paths:?}"
+    );
+
+    // Both b.rs and c.rs must still exist on the filesystem — no file
+    // removal is performed by orphans-only.
+    assert!(
+        b_path.exists(),
+        "b.rs must still be on disk after orphans-only reset"
+    );
+    assert!(
+        c_path.exists(),
+        "c.rs must still be on disk after orphans-only reset"
+    );
+
+    // Second reset must be idempotent: nothing left to purge.
+    let second = execute(ResetScope::OrphansOnly, &narrow_cfg)
+        .expect("second orphans-only reset must succeed");
+    assert_eq!(
+        second.orphans_purged, 0,
+        "second reset must be idempotent (orphans_purged == 0): {second:?}"
+    );
+    assert!(
+        second.purged_paths.is_empty(),
+        "second reset purged_paths must be empty: {:?}",
+        second.purged_paths
+    );
+}

crates/kebab-app/tests/twin_files_fetch_span.rs

@@ -0,0 +1,176 @@
+//! Regression test for the twin-file fetch_span media-type lookup bug.
+//!
+//! Twin files (identical content at different workspace paths) share one
+//! `assets` row whose PRIMARY KEY is the blake3 content hash. The old
+//! `fetch_span` implementation called
+//! `get_asset_by_workspace_path(&doc.workspace_path)` to check whether the
+//! media type was PDF/audio (and therefore reject span fetch). For a twin
+//! file that lookup could silently return the *other* twin's asset row if
+//! `assets.workspace_path` had been overwritten on the most recent ingest of
+//! the sibling — making the media-type branch decision incorrect.
+//!
+//! Fix: `fetch_span` now uses the 2-step lookup
+//!   `get_document_by_workspace_path` → `doc.source_asset_id` → `get_asset`
+//! so the result is always anchored to the requesting document, not
+//! whichever twin last updated `assets.workspace_path`.
+//!
+//! This test builds a twin-file scenario (two .md files at different paths
+//! with identical content), ingests both, then calls `fetch_span` on each
+//! twin's `doc_id` and asserts it succeeds. Before the fix, if the asset
+//! row's workspace_path happened to point at the wrong twin the span could
+//! return an incorrect `span_not_supported` for a non-PDF/audio file, or
+//! conversely allow span on a PDF twin by accident. After the fix, the
+//! lookup is always doc-specific.
+
+mod common;
+
+use common::TestEnv;
+use kebab_app::ingest_with_config;
+use kebab_core::{DocumentStore, FetchKind, FetchOpts, FetchQuery, IngestItemKind};
+
+#[test]
+fn twin_files_fetch_span_uses_correct_asset() {
+    let env = TestEnv::lexical_only();
+
+    // Write two markdown files with identical content at different paths.
+    let dir_a = env.workspace_root.join("src_a");
+    let dir_b = env.workspace_root.join("src_b");
+    std::fs::create_dir_all(&dir_a).unwrap();
+    std::fs::create_dir_all(&dir_b).unwrap();
+
+    // The content must produce at least 1 line so span fetch is non-trivial.
+    let content = "# Twin\n\nLine one.\n\nLine two.\n\nLine three.\n";
+    std::fs::write(dir_a.join("note.md"), content).unwrap();
+    std::fs::write(dir_b.join("note.md"), content).unwrap();
+
+    // Ingest all files (fixture workspace + our two new twins).
+    let report = ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("ingest must succeed");
+    assert_eq!(report.errors, 0, "no ingest errors; report={report:?}");
+
+    // Both twin paths must appear as New in the report.
+    let items = report.items.as_ref().expect("items must be present");
+    let twin_items: Vec<_> = items
+        .iter()
+        .filter(|i| {
+            i.doc_path.0.ends_with("src_a/note.md")
+                || i.doc_path.0.ends_with("src_b/note.md")
+        })
+        .collect();
+    assert_eq!(
+        twin_items.len(),
+        2,
+        "exactly 2 twin items expected; items={items:?}"
+    );
+    for item in &twin_items {
+        assert_eq!(
+            item.kind,
+            IngestItemKind::New,
+            "each twin must be New; item={item:?}"
+        );
+    }
+
+    // Resolve doc_ids for both workspace paths.
+    // The ingest layer normalises workspace_path to the path relative to
+    // workspace_root (e.g. "src_a/note.md"), so we look up by that form.
+    let store = kebab_store_sqlite::SqliteStore::open(&env.config).unwrap();
+    store.run_migrations().unwrap();
+
+    // Find the twin items by matching on suffix so the test is robust to
+    // however the workspace root is represented.
+    let items = report.items.as_ref().expect("items must be present");
+    let path_a_str = items
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("src_a/note.md"))
+        .map(|i| i.doc_path.0.clone())
+        .expect("src_a/note.md must appear in ingest report");
+    let path_b_str = items
+        .iter()
+        .find(|i| i.doc_path.0.ends_with("src_b/note.md"))
+        .map(|i| i.doc_path.0.clone())
+        .expect("src_b/note.md must appear in ingest report");
+
+    let path_a = kebab_core::WorkspacePath(path_a_str);
+    let path_b = kebab_core::WorkspacePath(path_b_str);
+
+    let doc_a = store
+        .get_document_by_workspace_path(&path_a)
+        .expect("get_document_by_workspace_path path_a")
+        .expect("doc_a must exist after ingest");
+    let doc_b = store
+        .get_document_by_workspace_path(&path_b)
+        .expect("get_document_by_workspace_path path_b")
+        .expect("doc_b must exist after ingest");
+
+    // Both twins share one asset_id (same content hash).
+    assert_eq!(
+        doc_a.source_asset_id, doc_b.source_asset_id,
+        "twin files must share one asset_id"
+    );
+
+    // Open App and issue span fetch on each twin's doc_id.
+    let app = env.app();
+
+    let result_a = app
+        .fetch(
+            FetchQuery::Span {
+                doc_id: doc_a.doc_id.clone(),
+                line_start: 1,
+                line_end: 2,
+            },
+            FetchOpts::default(),
+        )
+        .expect("fetch_span on twin A must succeed for a markdown file");
+    assert_eq!(result_a.kind, FetchKind::Span);
+    assert!(
+        result_a.text.as_deref().is_some_and(|t| !t.is_empty()),
+        "span text for twin A must not be empty"
+    );
+
+    let result_b = app
+        .fetch(
+            FetchQuery::Span {
+                doc_id: doc_b.doc_id.clone(),
+                line_start: 1,
+                line_end: 2,
+            },
+            FetchOpts::default(),
+        )
+        .expect("fetch_span on twin B must succeed for a markdown file");
+    assert_eq!(result_b.kind, FetchKind::Span);
+    assert!(
+        result_b.text.as_deref().is_some_and(|t| !t.is_empty()),
+        "span text for twin B must not be empty"
+    );
+
+    // Ingest again to force the asset.workspace_path flip-flop, then
+    // re-check. Pre-fix this was the scenario that triggered the bug:
+    // after the second ingest the asset row's workspace_path could point
+    // at either twin, making one twin's span fetch behave incorrectly.
+    let report2 = ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("second ingest must succeed");
+    assert_eq!(report2.errors, 0, "no ingest errors on second run; report={report2:?}");
+
+    // Re-open app after second ingest and verify span still works on both.
+    let app2 = env.app();
+
+    app2.fetch(
+        FetchQuery::Span {
+            doc_id: doc_a.doc_id.clone(),
+            line_start: 1,
+            line_end: 3,
+        },
+        FetchOpts::default(),
+    )
+    .expect("fetch_span on twin A after flip-flop must still succeed");
+
+    app2.fetch(
+        FetchQuery::Span {
+            doc_id: doc_b.doc_id.clone(),
+            line_start: 1,
+            line_end: 3,
+        },
+        FetchOpts::default(),
+    )
+    .expect("fetch_span on twin B after flip-flop must still succeed");
+}

crates/kebab-app/tests/twin_files_idempotent.rs

@@ -0,0 +1,90 @@
+//! Regression test for the twin-file idempotency bug.
+//!
+//! Identical-content files at different workspace paths share one
+//! `assets` row (`asset_id` = blake3 content hash, PRIMARY KEY). The
+//! old UPSERT `ON CONFLICT(asset_id) DO UPDATE SET workspace_path =
+//! excluded.workspace_path` made each twin overwrite the other's path
+//! on every ingest, so `get_asset_by_workspace_path(path1)` returned
+//! None (or the wrong twin) → re-process every time.
+//!
+//! Fix: `try_skip_unchanged` now uses `get_document_by_workspace_path`
+//! instead.  `documents.workspace_path` is UNIQUE (V001) so each twin
+//! has its own stable document row.
+//!
+//! Assertion contract:
+//!   1st ingest → 2 New (one per twin)
+//!   2nd ingest → 0 New, 0 Updated, 2 Unchanged
+
+mod common;
+
+use common::TestEnv;
+use kebab_app::ingest_with_config;
+use kebab_core::IngestItemKind;
+
+#[test]
+fn twin_files_second_ingest_is_unchanged() {
+    let env = TestEnv::lexical_only();
+
+    // Write two files with identical content at different paths.
+    let pkg_a = env.workspace_root.join("pkg_a");
+    let pkg_b = env.workspace_root.join("pkg_b");
+    std::fs::create_dir_all(&pkg_a).unwrap();
+    std::fs::create_dir_all(&pkg_b).unwrap();
+
+    let content = b"# shared\nThis content is identical in both files.\n";
+    std::fs::write(pkg_a.join("__init__.py"), content).unwrap();
+    std::fs::write(pkg_b.join("__init__.py"), content).unwrap();
+
+    // First ingest — both files must be New.
+    let first = ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("first ingest must succeed");
+    assert_eq!(first.errors, 0, "first ingest: no errors; report={first:?}");
+
+    let items = first.items.as_ref().expect("items must be present");
+    let twin_items: Vec<_> = items
+        .iter()
+        .filter(|i| {
+            i.doc_path.0.ends_with("__init__.py")
+        })
+        .collect();
+    assert_eq!(
+        twin_items.len(),
+        2,
+        "first ingest: expected exactly 2 __init__.py items; items={items:?}"
+    );
+    for item in &twin_items {
+        assert_eq!(
+            item.kind,
+            IngestItemKind::New,
+            "first ingest: each twin must be New; item={item:?}"
+        );
+    }
+
+    // Second ingest — same files, same content → both must be Unchanged.
+    let second = ingest_with_config(env.config.clone(), env.scope(), false)
+        .expect("second ingest must succeed");
+    assert_eq!(second.errors, 0, "second ingest: no errors; report={second:?}");
+    assert_eq!(second.new, 0, "second ingest: no new docs; report={second:?}");
+    assert_eq!(
+        second.updated, 0,
+        "second ingest: no updated docs (twin-file bug would set this to 2); report={second:?}"
+    );
+
+    let second_items = second.items.as_ref().expect("items must be present");
+    let twin_items2: Vec<_> = second_items
+        .iter()
+        .filter(|i| i.doc_path.0.ends_with("__init__.py"))
+        .collect();
+    assert_eq!(
+        twin_items2.len(),
+        2,
+        "second ingest: expected exactly 2 __init__.py items; items={second_items:?}"
+    );
+    for item in &twin_items2 {
+        assert_eq!(
+            item.kind,
+            IngestItemKind::Unchanged,
+            "second ingest: each twin must be Unchanged; item={item:?}"
+        );
+    }
+}

crates/kebab-chunk/src/code_go_ast_v1.rs

@@ -0,0 +1,322 @@
+//! `code-go-ast-v1` — maps a tree-sitter-derived Go AST
+//! `CanonicalDocument` (one `Block::Code` per semantic unit, each with
+//! `SourceSpan::Code`) to chunks 1:1. A unit longer than
+//! `AST_CHUNK_MAX_LINES` is split into `<symbol> [part i/N]` sub-chunks
+//! at blank-line paragraph boundaries (design §9.1 oversize fallback).
+//!
+//! tree-sitter is intentionally NOT a dependency here: AST work is
+//! parser-side (`kebab-parse-code`, design §6.3). This chunker only
+//! consumes the `CanonicalDocument`.
+//!
+//! `AST_CHUNK_MAX_LINES` is a constant matching
+//! `IngestCodeCfg::default().ast_chunk_max_lines` (200). Per-medium
+//! config threading needs a chunker registry (P+); same deviation
+//! pattern as `pdf-page-v1`'s pinned `chunker_version`
+//! (`tasks/HOTFIXES.md`).
+
+use kebab_core::{
+    Block, BlockId, CanonicalDocument, Chunk, ChunkPolicy, Chunker, ChunkerVersion, DocumentId,
+    SourceSpan, id_for_chunk,
+};
+
+const VERSION_LABEL: &str = "code-go-ast-v1";
+const BYTES_PER_TOKEN: usize = 3;
+const POLICY_HASH_HEX_LEN: usize = 16;
+const AST_CHUNK_MAX_LINES: u32 = 200;
+
+#[derive(Clone, Copy, Debug, Default)]
+pub struct CodeGoAstV1Chunker;
+
+impl Chunker for CodeGoAstV1Chunker {
+    fn chunker_version(&self) -> ChunkerVersion {
+        ChunkerVersion(VERSION_LABEL.to_string())
+    }
+
+    fn policy_hash(&self, policy: &ChunkPolicy) -> String {
+        let bytes = serde_json_canonicalizer::to_vec(policy)
+            .expect("canonical JSON serialization of ChunkPolicy must not fail");
+        let hex = blake3::hash(&bytes).to_hex().to_string();
+        hex[..POLICY_HASH_HEX_LEN].to_string()
+    }
+
+    fn chunk(
+        &self,
+        doc: &CanonicalDocument,
+        policy: &ChunkPolicy,
+    ) -> anyhow::Result<Vec<Chunk>> {
+        for b in &doc.blocks {
+            let c = match b {
+                Block::Code(c) => c,
+                _ => anyhow::bail!(
+                    "CodeGoAstV1Chunker only handles code docs (got non-Code block)"
+                ),
+            };
+            if !matches!(c.common.source_span, SourceSpan::Code { .. }) {
+                anyhow::bail!(
+                    "CodeGoAstV1Chunker only handles code docs (got non-Code source_span)"
+                );
+            }
+        }
+
+        let base_policy_hash = self.policy_hash(policy);
+        let chunker_version = self.chunker_version();
+        let mut out: Vec<Chunk> = Vec::new();
+
+        for b in &doc.blocks {
+            let cb = match b {
+                Block::Code(c) => c,
+                _ => unreachable!("validated above"),
+            };
+            let (ls, le, symbol, lang) = match &cb.common.source_span {
+                SourceSpan::Code { line_start, line_end, symbol, lang } => {
+                    (*line_start, *line_end, symbol.clone(), lang.clone())
+                }
+                _ => unreachable!("validated above"),
+            };
+            let block_ids: Vec<BlockId> = vec![cb.common.block_id.clone()];
+            let span_lines = le.saturating_sub(ls) + 1;
+
+            if span_lines <= AST_CHUNK_MAX_LINES {
+                let span = SourceSpan::Code {
+                    line_start: ls,
+                    line_end: le,
+                    symbol: symbol.clone(),
+                    lang: lang.clone(),
+                };
+                out.push(make_chunk(
+                    doc, &chunker_version, &block_ids, &base_policy_hash,
+                    None, span, cb.code.clone(),
+                ));
+            } else {
+                let parts = split_oversize(&cb.code);
+                let n = parts.len();
+                for (i, (off_start, off_end, text)) in parts.into_iter().enumerate() {
+                    let part_ls = ls + off_start;
+                    let part_le = ls + off_end;
+                    let part_sym = symbol
+                        .as_ref()
+                        .map(|s| format!("{s} [part {}/{n}]", i + 1));
+                    let span = SourceSpan::Code {
+                        line_start: part_ls,
+                        line_end: part_le,
+                        symbol: part_sym,
+                        lang: lang.clone(),
+                    };
+                    out.push(make_chunk(
+                        doc, &chunker_version, &block_ids, &base_policy_hash,
+                        Some(part_ls), span, text,
+                    ));
+                }
+            }
+        }
+
+        tracing::debug!(
+            target: "kebab-chunk",
+            doc_id = %doc.doc_id,
+            chunks = out.len(),
+            "code-go-ast-v1 chunked",
+        );
+        Ok(out)
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn make_chunk(
+    doc: &CanonicalDocument,
+    chunker_version: &ChunkerVersion,
+    block_ids: &[BlockId],
+    base_policy_hash: &str,
+    split_key: Option<u32>,
+    span: SourceSpan,
+    text: String,
+) -> Chunk {
+    let id_hash = match split_key {
+        Some(k) => format!("{base_policy_hash}#L{k}"),
+        None => base_policy_hash.to_string(),
+    };
+    let chunk_id = id_for_chunk(&doc.doc_id, chunker_version, block_ids, &id_hash);
+    let token_estimate = text.len().div_ceil(BYTES_PER_TOKEN);
+    Chunk {
+        chunk_id,
+        doc_id: DocumentId(doc.doc_id.0.clone()),
+        block_ids: block_ids.to_vec(),
+        text,
+        heading_path: Vec::new(),
+        source_spans: vec![span],
+        token_estimate,
+        chunker_version: chunker_version.clone(),
+        policy_hash: base_policy_hash.to_string(),
+    }
+}
+
+/// Split an oversize unit at blank-line paragraph boundaries, greedily
+/// gluing paragraphs until ~`AST_CHUNK_MAX_LINES` lines accumulate.
+/// Returns `(line_offset_start, line_offset_end, text)` where offsets are
+/// 0-based within the unit (caller adds the unit's absolute `line_start`).
+fn split_oversize(code: &str) -> Vec<(u32, u32, String)> {
+    let lines: Vec<&str> = code.split('\n').collect();
+    let total = lines.len() as u32;
+    let mut out: Vec<(u32, u32, String)> = Vec::new();
+    let mut start: u32 = 0;
+    while start < total {
+        let mut end = (start + AST_CHUNK_MAX_LINES).min(total);
+        let floor = start + (AST_CHUNK_MAX_LINES * 4 / 5);
+        if end < total {
+            if let Some(b) = (floor.min(end)..end)
+                .rev()
+                .find(|&i| lines[i as usize].trim().is_empty())
+            {
+                end = b + 1;
+            }
+        }
+        let text = lines[start as usize..end as usize].join("\n");
+        out.push((start, end.saturating_sub(1), text));
+        start = end;
+    }
+    if out.is_empty() {
+        out.push((0, total.saturating_sub(1), code.to_string()));
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{
+        Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+        SourceSpan, id_for_block, id_for_doc, AssetId, Lang, Metadata, ParserVersion, Provenance,
+        SourceType, TrustLevel, WorkspacePath,
+    };
+    use time::OffsetDateTime;
+
+    fn code_doc(units: &[(&str, u32, u32, &str)]) -> CanonicalDocument {
+        let wp = WorkspacePath("crates/x/src/a.go".into());
+        let aid = AssetId("a".repeat(64));
+        let pv = ParserVersion("code-go-v1".into());
+        let doc_id = id_for_doc(&wp, &aid, &pv);
+        let blocks = units
+            .iter()
+            .enumerate()
+            .map(|(i, (sym, ls, le, code))| {
+                let span = SourceSpan::Code {
+                    line_start: *ls,
+                    line_end: *le,
+                    symbol: Some((*sym).to_string()),
+                    lang: Some("go".into()),
+                };
+                let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+                Block::Code(CodeBlock {
+                    common: CommonBlock { block_id: bid, heading_path: vec![], source_span: span },
+                    lang: Some("go".into()),
+                    code: (*code).to_string(),
+                })
+            })
+            .collect();
+        CanonicalDocument {
+            doc_id, source_asset_id: aid, workspace_path: wp, title: "a".into(),
+            lang: Lang("und".into()), blocks,
+            metadata: Metadata {
+                aliases: vec![], tags: vec![],
+                created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                source_type: SourceType::Note, trust_level: TrustLevel::Primary,
+                user_id_alias: None, user: Default::default(),
+                repo: Some("kebab".into()), git_branch: Some("main".into()),
+                git_commit: Some("0".repeat(40)), code_lang: Some("go".into()),
+            },
+            provenance: Provenance { events: vec![] },
+            parser_version: pv, schema_version: 1, doc_version: 1,
+            last_chunker_version: None, last_embedding_version: None,
+        }
+    }
+    fn policy() -> ChunkPolicy {
+        ChunkPolicy { target_tokens: 500, overlap_tokens: 80,
+            respect_markdown_headings: false,
+            chunker_version: ChunkerVersion(VERSION_LABEL.into()) }
+    }
+
+    #[test]
+    fn chunker_version_is_code_go_ast_v1() {
+        assert_eq!(CodeGoAstV1Chunker.chunker_version(),
+            ChunkerVersion("code-go-ast-v1".into()));
+    }
+
+    #[test]
+    fn one_chunk_per_unit_preserves_code_span() {
+        let doc = code_doc(&[
+            ("parse", 1, 3, "func parse() {\n\t// x\n}"),
+            ("Foo.double", 5, 7, "func double() int {\n\t//\n\treturn 0\n}"),
+        ]);
+        let chunks = CodeGoAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert_eq!(chunks.len(), 2);
+        for c in &chunks {
+            assert_eq!(c.source_spans.len(), 1);
+            assert!(matches!(c.source_spans[0], SourceSpan::Code { .. }));
+            assert_eq!(c.heading_path, Vec::<String>::new());
+            assert_eq!(c.chunker_version.0, "code-go-ast-v1");
+        }
+        match &chunks[0].source_spans[0] {
+            SourceSpan::Code { symbol, line_start, line_end, .. } => {
+                assert_eq!(symbol.as_deref(), Some("parse"));
+                assert_eq!((*line_start, *line_end), (1, 3));
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    #[test]
+    fn oversize_unit_splits_into_parts_with_unique_ids() {
+        let body = (0..500).map(|i| format!("\tx{i} := {i}")).collect::<Vec<_>>().join("\n");
+        let code = format!("func big() {{\n{body}\n}}");
+        let doc = code_doc(&[("big", 1, 502, &code)]);
+        let chunks = CodeGoAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert!(chunks.len() >= 2, "oversize unit must split, got {}", chunks.len());
+        for c in &chunks {
+            match &c.source_spans[0] {
+                SourceSpan::Code { symbol, .. } => {
+                    assert!(symbol.as_deref().unwrap().starts_with("big [part "),
+                        "part-numbered symbol, got {symbol:?}");
+                }
+                _ => unreachable!(),
+            }
+        }
+        let mut ids: Vec<&str> = chunks.iter().map(|c| c.chunk_id.0.as_str()).collect();
+        let n = ids.len(); ids.sort(); ids.dedup();
+        assert_eq!(ids.len(), n, "chunk_ids unique across split parts");
+    }
+
+    #[test]
+    fn non_code_doc_errors() {
+        use kebab_core::TextBlock;
+        let mut doc = code_doc(&[("parse", 1, 1, "func parse() {}")]);
+        doc.blocks = vec![Block::Paragraph(TextBlock {
+            common: CommonBlock {
+                block_id: kebab_core::BlockId("b".into()),
+                heading_path: vec![],
+                source_span: SourceSpan::Line { start: 1, end: 1 },
+            },
+            text: "x".into(), inlines: vec![],
+        })];
+        let err = CodeGoAstV1Chunker.chunk(&doc, &policy()).unwrap_err();
+        assert!(err.to_string().contains("CodeGoAstV1Chunker"));
+    }
+
+    #[test]
+    fn deterministic_chunk_ids_1000() {
+        let doc = code_doc(&[("parse", 1, 2, "func parse() {}\n")]);
+        let base: Vec<String> = CodeGoAstV1Chunker.chunk(&doc, &policy())
+            .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+        for _ in 0..1000 {
+            let again: Vec<String> = CodeGoAstV1Chunker.chunk(&doc, &policy())
+                .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+            assert_eq!(again, base);
+        }
+    }
+
+    #[test]
+    fn policy_hash_matches_md_heading_v1() {
+        let p = policy();
+        assert_eq!(CodeGoAstV1Chunker.policy_hash(&p),
+            crate::MdHeadingV1Chunker.policy_hash(&p));
+    }
+}

crates/kebab-chunk/src/code_java_ast_v1.rs

@@ -0,0 +1,322 @@
+//! `code-java-ast-v1` — maps a tree-sitter-derived Java AST
+//! `CanonicalDocument` (one `Block::Code` per semantic unit, each with
+//! `SourceSpan::Code`) to chunks 1:1. A unit longer than
+//! `AST_CHUNK_MAX_LINES` is split into `<symbol> [part i/N]` sub-chunks
+//! at blank-line paragraph boundaries (design §9.1 oversize fallback).
+//!
+//! tree-sitter is intentionally NOT a dependency here: AST work is
+//! parser-side (`kebab-parse-code`, design §6.3). This chunker only
+//! consumes the `CanonicalDocument`.
+//!
+//! `AST_CHUNK_MAX_LINES` is a constant matching
+//! `IngestCodeCfg::default().ast_chunk_max_lines` (200). Per-medium
+//! config threading needs a chunker registry (P+); same deviation
+//! pattern as `pdf-page-v1`'s pinned `chunker_version`
+//! (`tasks/HOTFIXES.md`).
+
+use kebab_core::{
+    Block, BlockId, CanonicalDocument, Chunk, ChunkPolicy, Chunker, ChunkerVersion, DocumentId,
+    SourceSpan, id_for_chunk,
+};
+
+const VERSION_LABEL: &str = "code-java-ast-v1";
+const BYTES_PER_TOKEN: usize = 3;
+const POLICY_HASH_HEX_LEN: usize = 16;
+const AST_CHUNK_MAX_LINES: u32 = 200;
+
+#[derive(Clone, Copy, Debug, Default)]
+pub struct CodeJavaAstV1Chunker;
+
+impl Chunker for CodeJavaAstV1Chunker {
+    fn chunker_version(&self) -> ChunkerVersion {
+        ChunkerVersion(VERSION_LABEL.to_string())
+    }
+
+    fn policy_hash(&self, policy: &ChunkPolicy) -> String {
+        let bytes = serde_json_canonicalizer::to_vec(policy)
+            .expect("canonical JSON serialization of ChunkPolicy must not fail");
+        let hex = blake3::hash(&bytes).to_hex().to_string();
+        hex[..POLICY_HASH_HEX_LEN].to_string()
+    }
+
+    fn chunk(
+        &self,
+        doc: &CanonicalDocument,
+        policy: &ChunkPolicy,
+    ) -> anyhow::Result<Vec<Chunk>> {
+        for b in &doc.blocks {
+            let c = match b {
+                Block::Code(c) => c,
+                _ => anyhow::bail!(
+                    "CodeJavaAstV1Chunker only handles code docs (got non-Code block)"
+                ),
+            };
+            if !matches!(c.common.source_span, SourceSpan::Code { .. }) {
+                anyhow::bail!(
+                    "CodeJavaAstV1Chunker only handles code docs (got non-Code source_span)"
+                );
+            }
+        }
+
+        let base_policy_hash = self.policy_hash(policy);
+        let chunker_version = self.chunker_version();
+        let mut out: Vec<Chunk> = Vec::new();
+
+        for b in &doc.blocks {
+            let cb = match b {
+                Block::Code(c) => c,
+                _ => unreachable!("validated above"),
+            };
+            let (ls, le, symbol, lang) = match &cb.common.source_span {
+                SourceSpan::Code { line_start, line_end, symbol, lang } => {
+                    (*line_start, *line_end, symbol.clone(), lang.clone())
+                }
+                _ => unreachable!("validated above"),
+            };
+            let block_ids: Vec<BlockId> = vec![cb.common.block_id.clone()];
+            let span_lines = le.saturating_sub(ls) + 1;
+
+            if span_lines <= AST_CHUNK_MAX_LINES {
+                let span = SourceSpan::Code {
+                    line_start: ls,
+                    line_end: le,
+                    symbol: symbol.clone(),
+                    lang: lang.clone(),
+                };
+                out.push(make_chunk(
+                    doc, &chunker_version, &block_ids, &base_policy_hash,
+                    None, span, cb.code.clone(),
+                ));
+            } else {
+                let parts = split_oversize(&cb.code);
+                let n = parts.len();
+                for (i, (off_start, off_end, text)) in parts.into_iter().enumerate() {
+                    let part_ls = ls + off_start;
+                    let part_le = ls + off_end;
+                    let part_sym = symbol
+                        .as_ref()
+                        .map(|s| format!("{s} [part {}/{n}]", i + 1));
+                    let span = SourceSpan::Code {
+                        line_start: part_ls,
+                        line_end: part_le,
+                        symbol: part_sym,
+                        lang: lang.clone(),
+                    };
+                    out.push(make_chunk(
+                        doc, &chunker_version, &block_ids, &base_policy_hash,
+                        Some(part_ls), span, text,
+                    ));
+                }
+            }
+        }
+
+        tracing::debug!(
+            target: "kebab-chunk",
+            doc_id = %doc.doc_id,
+            chunks = out.len(),
+            "code-java-ast-v1 chunked",
+        );
+        Ok(out)
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn make_chunk(
+    doc: &CanonicalDocument,
+    chunker_version: &ChunkerVersion,
+    block_ids: &[BlockId],
+    base_policy_hash: &str,
+    split_key: Option<u32>,
+    span: SourceSpan,
+    text: String,
+) -> Chunk {
+    let id_hash = match split_key {
+        Some(k) => format!("{base_policy_hash}#L{k}"),
+        None => base_policy_hash.to_string(),
+    };
+    let chunk_id = id_for_chunk(&doc.doc_id, chunker_version, block_ids, &id_hash);
+    let token_estimate = text.len().div_ceil(BYTES_PER_TOKEN);
+    Chunk {
+        chunk_id,
+        doc_id: DocumentId(doc.doc_id.0.clone()),
+        block_ids: block_ids.to_vec(),
+        text,
+        heading_path: Vec::new(),
+        source_spans: vec![span],
+        token_estimate,
+        chunker_version: chunker_version.clone(),
+        policy_hash: base_policy_hash.to_string(),
+    }
+}
+
+/// Split an oversize unit at blank-line paragraph boundaries, greedily
+/// gluing paragraphs until ~`AST_CHUNK_MAX_LINES` lines accumulate.
+/// Returns `(line_offset_start, line_offset_end, text)` where offsets are
+/// 0-based within the unit (caller adds the unit's absolute `line_start`).
+fn split_oversize(code: &str) -> Vec<(u32, u32, String)> {
+    let lines: Vec<&str> = code.split('\n').collect();
+    let total = lines.len() as u32;
+    let mut out: Vec<(u32, u32, String)> = Vec::new();
+    let mut start: u32 = 0;
+    while start < total {
+        let mut end = (start + AST_CHUNK_MAX_LINES).min(total);
+        let floor = start + (AST_CHUNK_MAX_LINES * 4 / 5);
+        if end < total {
+            if let Some(b) = (floor.min(end)..end)
+                .rev()
+                .find(|&i| lines[i as usize].trim().is_empty())
+            {
+                end = b + 1;
+            }
+        }
+        let text = lines[start as usize..end as usize].join("\n");
+        out.push((start, end.saturating_sub(1), text));
+        start = end;
+    }
+    if out.is_empty() {
+        out.push((0, total.saturating_sub(1), code.to_string()));
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{
+        Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+        SourceSpan, id_for_block, id_for_doc, AssetId, Lang, Metadata, ParserVersion, Provenance,
+        SourceType, TrustLevel, WorkspacePath,
+    };
+    use time::OffsetDateTime;
+
+    fn code_doc(units: &[(&str, u32, u32, &str)]) -> CanonicalDocument {
+        let wp = WorkspacePath("crates/x/src/Main.java".into());
+        let aid = AssetId("a".repeat(64));
+        let pv = ParserVersion("code-java-v1".into());
+        let doc_id = id_for_doc(&wp, &aid, &pv);
+        let blocks = units
+            .iter()
+            .enumerate()
+            .map(|(i, (sym, ls, le, code))| {
+                let span = SourceSpan::Code {
+                    line_start: *ls,
+                    line_end: *le,
+                    symbol: Some((*sym).to_string()),
+                    lang: Some("java".into()),
+                };
+                let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+                Block::Code(CodeBlock {
+                    common: CommonBlock { block_id: bid, heading_path: vec![], source_span: span },
+                    lang: Some("java".into()),
+                    code: (*code).to_string(),
+                })
+            })
+            .collect();
+        CanonicalDocument {
+            doc_id, source_asset_id: aid, workspace_path: wp, title: "a".into(),
+            lang: Lang("und".into()), blocks,
+            metadata: Metadata {
+                aliases: vec![], tags: vec![],
+                created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                source_type: SourceType::Note, trust_level: TrustLevel::Primary,
+                user_id_alias: None, user: Default::default(),
+                repo: Some("kebab".into()), git_branch: Some("main".into()),
+                git_commit: Some("0".repeat(40)), code_lang: Some("java".into()),
+            },
+            provenance: Provenance { events: vec![] },
+            parser_version: pv, schema_version: 1, doc_version: 1,
+            last_chunker_version: None, last_embedding_version: None,
+        }
+    }
+    fn policy() -> ChunkPolicy {
+        ChunkPolicy { target_tokens: 500, overlap_tokens: 80,
+            respect_markdown_headings: false,
+            chunker_version: ChunkerVersion(VERSION_LABEL.into()) }
+    }
+
+    #[test]
+    fn chunker_version_is_code_java_ast_v1() {
+        assert_eq!(CodeJavaAstV1Chunker.chunker_version(),
+            ChunkerVersion("code-java-ast-v1".into()));
+    }
+
+    #[test]
+    fn one_chunk_per_unit_preserves_code_span() {
+        let doc = code_doc(&[
+            ("parse", 1, 3, "void parse() {\n\t// x\n}"),
+            ("Foo.double", 5, 7, "int double() {\n\t//\n\treturn 0;\n}"),
+        ]);
+        let chunks = CodeJavaAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert_eq!(chunks.len(), 2);
+        for c in &chunks {
+            assert_eq!(c.source_spans.len(), 1);
+            assert!(matches!(c.source_spans[0], SourceSpan::Code { .. }));
+            assert_eq!(c.heading_path, Vec::<String>::new());
+            assert_eq!(c.chunker_version.0, "code-java-ast-v1");
+        }
+        match &chunks[0].source_spans[0] {
+            SourceSpan::Code { symbol, line_start, line_end, .. } => {
+                assert_eq!(symbol.as_deref(), Some("parse"));
+                assert_eq!((*line_start, *line_end), (1, 3));
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    #[test]
+    fn oversize_unit_splits_into_parts_with_unique_ids() {
+        let body = (0..500).map(|i| format!("\tint x{i} = {i};")).collect::<Vec<_>>().join("\n");
+        let code = format!("void big() {{\n{body}\n}}");
+        let doc = code_doc(&[("big", 1, 502, &code)]);
+        let chunks = CodeJavaAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert!(chunks.len() >= 2, "oversize unit must split, got {}", chunks.len());
+        for c in &chunks {
+            match &c.source_spans[0] {
+                SourceSpan::Code { symbol, .. } => {
+                    assert!(symbol.as_deref().unwrap().starts_with("big [part "),
+                        "part-numbered symbol, got {symbol:?}");
+                }
+                _ => unreachable!(),
+            }
+        }
+        let mut ids: Vec<&str> = chunks.iter().map(|c| c.chunk_id.0.as_str()).collect();
+        let n = ids.len(); ids.sort(); ids.dedup();
+        assert_eq!(ids.len(), n, "chunk_ids unique across split parts");
+    }
+
+    #[test]
+    fn non_code_doc_errors() {
+        use kebab_core::TextBlock;
+        let mut doc = code_doc(&[("parse", 1, 1, "void parse() {}")]);
+        doc.blocks = vec![Block::Paragraph(TextBlock {
+            common: CommonBlock {
+                block_id: kebab_core::BlockId("b".into()),
+                heading_path: vec![],
+                source_span: SourceSpan::Line { start: 1, end: 1 },
+            },
+            text: "x".into(), inlines: vec![],
+        })];
+        let err = CodeJavaAstV1Chunker.chunk(&doc, &policy()).unwrap_err();
+        assert!(err.to_string().contains("CodeJavaAstV1Chunker"));
+    }
+
+    #[test]
+    fn deterministic_chunk_ids_1000() {
+        let doc = code_doc(&[("parse", 1, 2, "void parse() {}\n")]);
+        let base: Vec<String> = CodeJavaAstV1Chunker.chunk(&doc, &policy())
+            .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+        for _ in 0..1000 {
+            let again: Vec<String> = CodeJavaAstV1Chunker.chunk(&doc, &policy())
+                .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+            assert_eq!(again, base);
+        }
+    }
+
+    #[test]
+    fn policy_hash_matches_md_heading_v1() {
+        let p = policy();
+        assert_eq!(CodeJavaAstV1Chunker.policy_hash(&p),
+            crate::MdHeadingV1Chunker.policy_hash(&p));
+    }
+}

crates/kebab-chunk/src/code_js_ast_v1.rs

@@ -0,0 +1,322 @@
+//! `code-js-ast-v1` — maps a tree-sitter-derived JavaScript AST
+//! `CanonicalDocument` (one `Block::Code` per semantic unit, each with
+//! `SourceSpan::Code`) to chunks 1:1. A unit longer than
+//! `AST_CHUNK_MAX_LINES` is split into `<symbol> [part i/N]` sub-chunks
+//! at blank-line paragraph boundaries (design §9.1 oversize fallback).
+//!
+//! tree-sitter is intentionally NOT a dependency here: AST work is
+//! parser-side (`kebab-parse-code`, design §6.3). This chunker only
+//! consumes the `CanonicalDocument`.
+//!
+//! `AST_CHUNK_MAX_LINES` is a constant matching
+//! `IngestCodeCfg::default().ast_chunk_max_lines` (200). Per-medium
+//! config threading needs a chunker registry (P+); same deviation
+//! pattern as `pdf-page-v1`'s pinned `chunker_version`
+//! (`tasks/HOTFIXES.md`).
+
+use kebab_core::{
+    Block, BlockId, CanonicalDocument, Chunk, ChunkPolicy, Chunker, ChunkerVersion, DocumentId,
+    SourceSpan, id_for_chunk,
+};
+
+const VERSION_LABEL: &str = "code-js-ast-v1";
+const BYTES_PER_TOKEN: usize = 3;
+const POLICY_HASH_HEX_LEN: usize = 16;
+const AST_CHUNK_MAX_LINES: u32 = 200;
+
+#[derive(Clone, Copy, Debug, Default)]
+pub struct CodeJsAstV1Chunker;
+
+impl Chunker for CodeJsAstV1Chunker {
+    fn chunker_version(&self) -> ChunkerVersion {
+        ChunkerVersion(VERSION_LABEL.to_string())
+    }
+
+    fn policy_hash(&self, policy: &ChunkPolicy) -> String {
+        let bytes = serde_json_canonicalizer::to_vec(policy)
+            .expect("canonical JSON serialization of ChunkPolicy must not fail");
+        let hex = blake3::hash(&bytes).to_hex().to_string();
+        hex[..POLICY_HASH_HEX_LEN].to_string()
+    }
+
+    fn chunk(
+        &self,
+        doc: &CanonicalDocument,
+        policy: &ChunkPolicy,
+    ) -> anyhow::Result<Vec<Chunk>> {
+        for b in &doc.blocks {
+            let c = match b {
+                Block::Code(c) => c,
+                _ => anyhow::bail!(
+                    "CodeJsAstV1Chunker only handles code docs (got non-Code block)"
+                ),
+            };
+            if !matches!(c.common.source_span, SourceSpan::Code { .. }) {
+                anyhow::bail!(
+                    "CodeJsAstV1Chunker only handles code docs (got non-Code source_span)"
+                );
+            }
+        }
+
+        let base_policy_hash = self.policy_hash(policy);
+        let chunker_version = self.chunker_version();
+        let mut out: Vec<Chunk> = Vec::new();
+
+        for b in &doc.blocks {
+            let cb = match b {
+                Block::Code(c) => c,
+                _ => unreachable!("validated above"),
+            };
+            let (ls, le, symbol, lang) = match &cb.common.source_span {
+                SourceSpan::Code { line_start, line_end, symbol, lang } => {
+                    (*line_start, *line_end, symbol.clone(), lang.clone())
+                }
+                _ => unreachable!("validated above"),
+            };
+            let block_ids: Vec<BlockId> = vec![cb.common.block_id.clone()];
+            let span_lines = le.saturating_sub(ls) + 1;
+
+            if span_lines <= AST_CHUNK_MAX_LINES {
+                let span = SourceSpan::Code {
+                    line_start: ls,
+                    line_end: le,
+                    symbol: symbol.clone(),
+                    lang: lang.clone(),
+                };
+                out.push(make_chunk(
+                    doc, &chunker_version, &block_ids, &base_policy_hash,
+                    None, span, cb.code.clone(),
+                ));
+            } else {
+                let parts = split_oversize(&cb.code);
+                let n = parts.len();
+                for (i, (off_start, off_end, text)) in parts.into_iter().enumerate() {
+                    let part_ls = ls + off_start;
+                    let part_le = ls + off_end;
+                    let part_sym = symbol
+                        .as_ref()
+                        .map(|s| format!("{s} [part {}/{n}]", i + 1));
+                    let span = SourceSpan::Code {
+                        line_start: part_ls,
+                        line_end: part_le,
+                        symbol: part_sym,
+                        lang: lang.clone(),
+                    };
+                    out.push(make_chunk(
+                        doc, &chunker_version, &block_ids, &base_policy_hash,
+                        Some(part_ls), span, text,
+                    ));
+                }
+            }
+        }
+
+        tracing::debug!(
+            target: "kebab-chunk",
+            doc_id = %doc.doc_id,
+            chunks = out.len(),
+            "code-js-ast-v1 chunked",
+        );
+        Ok(out)
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn make_chunk(
+    doc: &CanonicalDocument,
+    chunker_version: &ChunkerVersion,
+    block_ids: &[BlockId],
+    base_policy_hash: &str,
+    split_key: Option<u32>,
+    span: SourceSpan,
+    text: String,
+) -> Chunk {
+    let id_hash = match split_key {
+        Some(k) => format!("{base_policy_hash}#L{k}"),
+        None => base_policy_hash.to_string(),
+    };
+    let chunk_id = id_for_chunk(&doc.doc_id, chunker_version, block_ids, &id_hash);
+    let token_estimate = text.len().div_ceil(BYTES_PER_TOKEN);
+    Chunk {
+        chunk_id,
+        doc_id: DocumentId(doc.doc_id.0.clone()),
+        block_ids: block_ids.to_vec(),
+        text,
+        heading_path: Vec::new(),
+        source_spans: vec![span],
+        token_estimate,
+        chunker_version: chunker_version.clone(),
+        policy_hash: base_policy_hash.to_string(),
+    }
+}
+
+/// Split an oversize unit at blank-line paragraph boundaries, greedily
+/// gluing paragraphs until ~`AST_CHUNK_MAX_LINES` lines accumulate.
+/// Returns `(line_offset_start, line_offset_end, text)` where offsets are
+/// 0-based within the unit (caller adds the unit's absolute `line_start`).
+fn split_oversize(code: &str) -> Vec<(u32, u32, String)> {
+    let lines: Vec<&str> = code.split('\n').collect();
+    let total = lines.len() as u32;
+    let mut out: Vec<(u32, u32, String)> = Vec::new();
+    let mut start: u32 = 0;
+    while start < total {
+        let mut end = (start + AST_CHUNK_MAX_LINES).min(total);
+        let floor = start + (AST_CHUNK_MAX_LINES * 4 / 5);
+        if end < total {
+            if let Some(b) = (floor.min(end)..end)
+                .rev()
+                .find(|&i| lines[i as usize].trim().is_empty())
+            {
+                end = b + 1;
+            }
+        }
+        let text = lines[start as usize..end as usize].join("\n");
+        out.push((start, end.saturating_sub(1), text));
+        start = end;
+    }
+    if out.is_empty() {
+        out.push((0, total.saturating_sub(1), code.to_string()));
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{
+        Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+        SourceSpan, id_for_block, id_for_doc, AssetId, Lang, Metadata, ParserVersion, Provenance,
+        SourceType, TrustLevel, WorkspacePath,
+    };
+    use time::OffsetDateTime;
+
+    fn code_doc(units: &[(&str, u32, u32, &str)]) -> CanonicalDocument {
+        let wp = WorkspacePath("crates/x/src/a.js".into());
+        let aid = AssetId("a".repeat(64));
+        let pv = ParserVersion("code-js-v1".into());
+        let doc_id = id_for_doc(&wp, &aid, &pv);
+        let blocks = units
+            .iter()
+            .enumerate()
+            .map(|(i, (sym, ls, le, code))| {
+                let span = SourceSpan::Code {
+                    line_start: *ls,
+                    line_end: *le,
+                    symbol: Some((*sym).to_string()),
+                    lang: Some("javascript".into()),
+                };
+                let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+                Block::Code(CodeBlock {
+                    common: CommonBlock { block_id: bid, heading_path: vec![], source_span: span },
+                    lang: Some("javascript".into()),
+                    code: (*code).to_string(),
+                })
+            })
+            .collect();
+        CanonicalDocument {
+            doc_id, source_asset_id: aid, workspace_path: wp, title: "a".into(),
+            lang: Lang("und".into()), blocks,
+            metadata: Metadata {
+                aliases: vec![], tags: vec![],
+                created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                source_type: SourceType::Note, trust_level: TrustLevel::Primary,
+                user_id_alias: None, user: Default::default(),
+                repo: Some("kebab".into()), git_branch: Some("main".into()),
+                git_commit: Some("0".repeat(40)), code_lang: Some("javascript".into()),
+            },
+            provenance: Provenance { events: vec![] },
+            parser_version: pv, schema_version: 1, doc_version: 1,
+            last_chunker_version: None, last_embedding_version: None,
+        }
+    }
+    fn policy() -> ChunkPolicy {
+        ChunkPolicy { target_tokens: 500, overlap_tokens: 80,
+            respect_markdown_headings: false,
+            chunker_version: ChunkerVersion(VERSION_LABEL.into()) }
+    }
+
+    #[test]
+    fn chunker_version_is_code_js_ast_v1() {
+        assert_eq!(CodeJsAstV1Chunker.chunker_version(),
+            ChunkerVersion("code-js-ast-v1".into()));
+    }
+
+    #[test]
+    fn one_chunk_per_unit_preserves_code_span() {
+        let doc = code_doc(&[
+            ("parse", 1, 3, "function parse() {\n    // x\n}"),
+            ("Foo.double", 5, 7, "function double() {\n    //\n    return 0;\n}"),
+        ]);
+        let chunks = CodeJsAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert_eq!(chunks.len(), 2);
+        for c in &chunks {
+            assert_eq!(c.source_spans.len(), 1);
+            assert!(matches!(c.source_spans[0], SourceSpan::Code { .. }));
+            assert_eq!(c.heading_path, Vec::<String>::new());
+            assert_eq!(c.chunker_version.0, "code-js-ast-v1");
+        }
+        match &chunks[0].source_spans[0] {
+            SourceSpan::Code { symbol, line_start, line_end, .. } => {
+                assert_eq!(symbol.as_deref(), Some("parse"));
+                assert_eq!((*line_start, *line_end), (1, 3));
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    #[test]
+    fn oversize_unit_splits_into_parts_with_unique_ids() {
+        let body = (0..500).map(|i| format!("    const x{i} = {i};")).collect::<Vec<_>>().join("\n");
+        let code = format!("function big() {{\n{body}\n}}");
+        let doc = code_doc(&[("big", 1, 502, &code)]);
+        let chunks = CodeJsAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert!(chunks.len() >= 2, "oversize unit must split, got {}", chunks.len());
+        for c in &chunks {
+            match &c.source_spans[0] {
+                SourceSpan::Code { symbol, .. } => {
+                    assert!(symbol.as_deref().unwrap().starts_with("big [part "),
+                        "part-numbered symbol, got {symbol:?}");
+                }
+                _ => unreachable!(),
+            }
+        }
+        let mut ids: Vec<&str> = chunks.iter().map(|c| c.chunk_id.0.as_str()).collect();
+        let n = ids.len(); ids.sort(); ids.dedup();
+        assert_eq!(ids.len(), n, "chunk_ids unique across split parts");
+    }
+
+    #[test]
+    fn non_code_doc_errors() {
+        use kebab_core::TextBlock;
+        let mut doc = code_doc(&[("parse", 1, 1, "function parse() {}")]);
+        doc.blocks = vec![Block::Paragraph(TextBlock {
+            common: CommonBlock {
+                block_id: kebab_core::BlockId("b".into()),
+                heading_path: vec![],
+                source_span: SourceSpan::Line { start: 1, end: 1 },
+            },
+            text: "x".into(), inlines: vec![],
+        })];
+        let err = CodeJsAstV1Chunker.chunk(&doc, &policy()).unwrap_err();
+        assert!(err.to_string().contains("CodeJsAstV1Chunker"));
+    }
+
+    #[test]
+    fn deterministic_chunk_ids_1000() {
+        let doc = code_doc(&[("parse", 1, 2, "function parse() {}\n")]);
+        let base: Vec<String> = CodeJsAstV1Chunker.chunk(&doc, &policy())
+            .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+        for _ in 0..1000 {
+            let again: Vec<String> = CodeJsAstV1Chunker.chunk(&doc, &policy())
+                .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+            assert_eq!(again, base);
+        }
+    }
+
+    #[test]
+    fn policy_hash_matches_md_heading_v1() {
+        let p = policy();
+        assert_eq!(CodeJsAstV1Chunker.policy_hash(&p),
+            crate::MdHeadingV1Chunker.policy_hash(&p));
+    }
+}

crates/kebab-chunk/src/code_kotlin_ast_v1.rs

@@ -0,0 +1,322 @@
+//! `code-kotlin-ast-v1` — maps a tree-sitter-derived Kotlin AST
+//! `CanonicalDocument` (one `Block::Code` per semantic unit, each with
+//! `SourceSpan::Code`) to chunks 1:1. A unit longer than
+//! `AST_CHUNK_MAX_LINES` is split into `<symbol> [part i/N]` sub-chunks
+//! at blank-line paragraph boundaries (design §9.1 oversize fallback).
+//!
+//! tree-sitter is intentionally NOT a dependency here: AST work is
+//! parser-side (`kebab-parse-code`, design §6.3). This chunker only
+//! consumes the `CanonicalDocument`.
+//!
+//! `AST_CHUNK_MAX_LINES` is a constant matching
+//! `IngestCodeCfg::default().ast_chunk_max_lines` (200). Per-medium
+//! config threading needs a chunker registry (P+); same deviation
+//! pattern as `pdf-page-v1`'s pinned `chunker_version`
+//! (`tasks/HOTFIXES.md`).
+
+use kebab_core::{
+    Block, BlockId, CanonicalDocument, Chunk, ChunkPolicy, Chunker, ChunkerVersion, DocumentId,
+    SourceSpan, id_for_chunk,
+};
+
+const VERSION_LABEL: &str = "code-kotlin-ast-v1";
+const BYTES_PER_TOKEN: usize = 3;
+const POLICY_HASH_HEX_LEN: usize = 16;
+const AST_CHUNK_MAX_LINES: u32 = 200;
+
+#[derive(Clone, Copy, Debug, Default)]
+pub struct CodeKotlinAstV1Chunker;
+
+impl Chunker for CodeKotlinAstV1Chunker {
+    fn chunker_version(&self) -> ChunkerVersion {
+        ChunkerVersion(VERSION_LABEL.to_string())
+    }
+
+    fn policy_hash(&self, policy: &ChunkPolicy) -> String {
+        let bytes = serde_json_canonicalizer::to_vec(policy)
+            .expect("canonical JSON serialization of ChunkPolicy must not fail");
+        let hex = blake3::hash(&bytes).to_hex().to_string();
+        hex[..POLICY_HASH_HEX_LEN].to_string()
+    }
+
+    fn chunk(
+        &self,
+        doc: &CanonicalDocument,
+        policy: &ChunkPolicy,
+    ) -> anyhow::Result<Vec<Chunk>> {
+        for b in &doc.blocks {
+            let c = match b {
+                Block::Code(c) => c,
+                _ => anyhow::bail!(
+                    "CodeKotlinAstV1Chunker only handles code docs (got non-Code block)"
+                ),
+            };
+            if !matches!(c.common.source_span, SourceSpan::Code { .. }) {
+                anyhow::bail!(
+                    "CodeKotlinAstV1Chunker only handles code docs (got non-Code source_span)"
+                );
+            }
+        }
+
+        let base_policy_hash = self.policy_hash(policy);
+        let chunker_version = self.chunker_version();
+        let mut out: Vec<Chunk> = Vec::new();
+
+        for b in &doc.blocks {
+            let cb = match b {
+                Block::Code(c) => c,
+                _ => unreachable!("validated above"),
+            };
+            let (ls, le, symbol, lang) = match &cb.common.source_span {
+                SourceSpan::Code { line_start, line_end, symbol, lang } => {
+                    (*line_start, *line_end, symbol.clone(), lang.clone())
+                }
+                _ => unreachable!("validated above"),
+            };
+            let block_ids: Vec<BlockId> = vec![cb.common.block_id.clone()];
+            let span_lines = le.saturating_sub(ls) + 1;
+
+            if span_lines <= AST_CHUNK_MAX_LINES {
+                let span = SourceSpan::Code {
+                    line_start: ls,
+                    line_end: le,
+                    symbol: symbol.clone(),
+                    lang: lang.clone(),
+                };
+                out.push(make_chunk(
+                    doc, &chunker_version, &block_ids, &base_policy_hash,
+                    None, span, cb.code.clone(),
+                ));
+            } else {
+                let parts = split_oversize(&cb.code);
+                let n = parts.len();
+                for (i, (off_start, off_end, text)) in parts.into_iter().enumerate() {
+                    let part_ls = ls + off_start;
+                    let part_le = ls + off_end;
+                    let part_sym = symbol
+                        .as_ref()
+                        .map(|s| format!("{s} [part {}/{n}]", i + 1));
+                    let span = SourceSpan::Code {
+                        line_start: part_ls,
+                        line_end: part_le,
+                        symbol: part_sym,
+                        lang: lang.clone(),
+                    };
+                    out.push(make_chunk(
+                        doc, &chunker_version, &block_ids, &base_policy_hash,
+                        Some(part_ls), span, text,
+                    ));
+                }
+            }
+        }
+
+        tracing::debug!(
+            target: "kebab-chunk",
+            doc_id = %doc.doc_id,
+            chunks = out.len(),
+            "code-kotlin-ast-v1 chunked",
+        );
+        Ok(out)
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn make_chunk(
+    doc: &CanonicalDocument,
+    chunker_version: &ChunkerVersion,
+    block_ids: &[BlockId],
+    base_policy_hash: &str,
+    split_key: Option<u32>,
+    span: SourceSpan,
+    text: String,
+) -> Chunk {
+    let id_hash = match split_key {
+        Some(k) => format!("{base_policy_hash}#L{k}"),
+        None => base_policy_hash.to_string(),
+    };
+    let chunk_id = id_for_chunk(&doc.doc_id, chunker_version, block_ids, &id_hash);
+    let token_estimate = text.len().div_ceil(BYTES_PER_TOKEN);
+    Chunk {
+        chunk_id,
+        doc_id: DocumentId(doc.doc_id.0.clone()),
+        block_ids: block_ids.to_vec(),
+        text,
+        heading_path: Vec::new(),
+        source_spans: vec![span],
+        token_estimate,
+        chunker_version: chunker_version.clone(),
+        policy_hash: base_policy_hash.to_string(),
+    }
+}
+
+/// Split an oversize unit at blank-line paragraph boundaries, greedily
+/// gluing paragraphs until ~`AST_CHUNK_MAX_LINES` lines accumulate.
+/// Returns `(line_offset_start, line_offset_end, text)` where offsets are
+/// 0-based within the unit (caller adds the unit's absolute `line_start`).
+fn split_oversize(code: &str) -> Vec<(u32, u32, String)> {
+    let lines: Vec<&str> = code.split('\n').collect();
+    let total = lines.len() as u32;
+    let mut out: Vec<(u32, u32, String)> = Vec::new();
+    let mut start: u32 = 0;
+    while start < total {
+        let mut end = (start + AST_CHUNK_MAX_LINES).min(total);
+        let floor = start + (AST_CHUNK_MAX_LINES * 4 / 5);
+        if end < total {
+            if let Some(b) = (floor.min(end)..end)
+                .rev()
+                .find(|&i| lines[i as usize].trim().is_empty())
+            {
+                end = b + 1;
+            }
+        }
+        let text = lines[start as usize..end as usize].join("\n");
+        out.push((start, end.saturating_sub(1), text));
+        start = end;
+    }
+    if out.is_empty() {
+        out.push((0, total.saturating_sub(1), code.to_string()));
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{
+        Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+        SourceSpan, id_for_block, id_for_doc, AssetId, Lang, Metadata, ParserVersion, Provenance,
+        SourceType, TrustLevel, WorkspacePath,
+    };
+    use time::OffsetDateTime;
+
+    fn code_doc(units: &[(&str, u32, u32, &str)]) -> CanonicalDocument {
+        let wp = WorkspacePath("crates/x/src/Main.kt".into());
+        let aid = AssetId("a".repeat(64));
+        let pv = ParserVersion("code-kotlin-v1".into());
+        let doc_id = id_for_doc(&wp, &aid, &pv);
+        let blocks = units
+            .iter()
+            .enumerate()
+            .map(|(i, (sym, ls, le, code))| {
+                let span = SourceSpan::Code {
+                    line_start: *ls,
+                    line_end: *le,
+                    symbol: Some((*sym).to_string()),
+                    lang: Some("kotlin".into()),
+                };
+                let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+                Block::Code(CodeBlock {
+                    common: CommonBlock { block_id: bid, heading_path: vec![], source_span: span },
+                    lang: Some("kotlin".into()),
+                    code: (*code).to_string(),
+                })
+            })
+            .collect();
+        CanonicalDocument {
+            doc_id, source_asset_id: aid, workspace_path: wp, title: "a".into(),
+            lang: Lang("und".into()), blocks,
+            metadata: Metadata {
+                aliases: vec![], tags: vec![],
+                created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                source_type: SourceType::Note, trust_level: TrustLevel::Primary,
+                user_id_alias: None, user: Default::default(),
+                repo: Some("kebab".into()), git_branch: Some("main".into()),
+                git_commit: Some("0".repeat(40)), code_lang: Some("kotlin".into()),
+            },
+            provenance: Provenance { events: vec![] },
+            parser_version: pv, schema_version: 1, doc_version: 1,
+            last_chunker_version: None, last_embedding_version: None,
+        }
+    }
+    fn policy() -> ChunkPolicy {
+        ChunkPolicy { target_tokens: 500, overlap_tokens: 80,
+            respect_markdown_headings: false,
+            chunker_version: ChunkerVersion(VERSION_LABEL.into()) }
+    }
+
+    #[test]
+    fn chunker_version_is_code_kotlin_ast_v1() {
+        assert_eq!(CodeKotlinAstV1Chunker.chunker_version(),
+            ChunkerVersion("code-kotlin-ast-v1".into()));
+    }
+
+    #[test]
+    fn one_chunk_per_unit_preserves_code_span() {
+        let doc = code_doc(&[
+            ("parse", 1, 3, "fun parse() {\n\t// x\n}"),
+            ("Foo.double", 5, 7, "fun double(): Int {\n\t//\n\treturn 0\n}"),
+        ]);
+        let chunks = CodeKotlinAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert_eq!(chunks.len(), 2);
+        for c in &chunks {
+            assert_eq!(c.source_spans.len(), 1);
+            assert!(matches!(c.source_spans[0], SourceSpan::Code { .. }));
+            assert_eq!(c.heading_path, Vec::<String>::new());
+            assert_eq!(c.chunker_version.0, "code-kotlin-ast-v1");
+        }
+        match &chunks[0].source_spans[0] {
+            SourceSpan::Code { symbol, line_start, line_end, .. } => {
+                assert_eq!(symbol.as_deref(), Some("parse"));
+                assert_eq!((*line_start, *line_end), (1, 3));
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    #[test]
+    fn oversize_unit_splits_into_parts_with_unique_ids() {
+        let body = (0..500).map(|i| format!("\tval x{i} = {i}")).collect::<Vec<_>>().join("\n");
+        let code = format!("fun big() {{\n{body}\n}}");
+        let doc = code_doc(&[("big", 1, 502, &code)]);
+        let chunks = CodeKotlinAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert!(chunks.len() >= 2, "oversize unit must split, got {}", chunks.len());
+        for c in &chunks {
+            match &c.source_spans[0] {
+                SourceSpan::Code { symbol, .. } => {
+                    assert!(symbol.as_deref().unwrap().starts_with("big [part "),
+                        "part-numbered symbol, got {symbol:?}");
+                }
+                _ => unreachable!(),
+            }
+        }
+        let mut ids: Vec<&str> = chunks.iter().map(|c| c.chunk_id.0.as_str()).collect();
+        let n = ids.len(); ids.sort(); ids.dedup();
+        assert_eq!(ids.len(), n, "chunk_ids unique across split parts");
+    }
+
+    #[test]
+    fn non_code_doc_errors() {
+        use kebab_core::TextBlock;
+        let mut doc = code_doc(&[("parse", 1, 1, "fun parse() {}")]);
+        doc.blocks = vec![Block::Paragraph(TextBlock {
+            common: CommonBlock {
+                block_id: kebab_core::BlockId("b".into()),
+                heading_path: vec![],
+                source_span: SourceSpan::Line { start: 1, end: 1 },
+            },
+            text: "x".into(), inlines: vec![],
+        })];
+        let err = CodeKotlinAstV1Chunker.chunk(&doc, &policy()).unwrap_err();
+        assert!(err.to_string().contains("CodeKotlinAstV1Chunker"));
+    }
+
+    #[test]
+    fn deterministic_chunk_ids_1000() {
+        let doc = code_doc(&[("parse", 1, 2, "fun parse() {}\n")]);
+        let base: Vec<String> = CodeKotlinAstV1Chunker.chunk(&doc, &policy())
+            .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+        for _ in 0..1000 {
+            let again: Vec<String> = CodeKotlinAstV1Chunker.chunk(&doc, &policy())
+                .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+            assert_eq!(again, base);
+        }
+    }
+
+    #[test]
+    fn policy_hash_matches_md_heading_v1() {
+        let p = policy();
+        assert_eq!(CodeKotlinAstV1Chunker.policy_hash(&p),
+            crate::MdHeadingV1Chunker.policy_hash(&p));
+    }
+}

crates/kebab-chunk/src/code_python_ast_v1.rs

@@ -0,0 +1,322 @@
+//! `code-python-ast-v1` — maps a tree-sitter-derived Python AST
+//! `CanonicalDocument` (one `Block::Code` per semantic unit, each with
+//! `SourceSpan::Code`) to chunks 1:1. A unit longer than
+//! `AST_CHUNK_MAX_LINES` is split into `<symbol> [part i/N]` sub-chunks
+//! at blank-line paragraph boundaries (design §9.1 oversize fallback).
+//!
+//! tree-sitter is intentionally NOT a dependency here: AST work is
+//! parser-side (`kebab-parse-code`, design §6.3). This chunker only
+//! consumes the `CanonicalDocument`.
+//!
+//! `AST_CHUNK_MAX_LINES` is a constant matching
+//! `IngestCodeCfg::default().ast_chunk_max_lines` (200). Per-medium
+//! config threading needs a chunker registry (P+); same deviation
+//! pattern as `pdf-page-v1`'s pinned `chunker_version`
+//! (`tasks/HOTFIXES.md`).
+
+use kebab_core::{
+    Block, BlockId, CanonicalDocument, Chunk, ChunkPolicy, Chunker, ChunkerVersion, DocumentId,
+    SourceSpan, id_for_chunk,
+};
+
+const VERSION_LABEL: &str = "code-python-ast-v1";
+const BYTES_PER_TOKEN: usize = 3;
+const POLICY_HASH_HEX_LEN: usize = 16;
+const AST_CHUNK_MAX_LINES: u32 = 200;
+
+#[derive(Clone, Copy, Debug, Default)]
+pub struct CodePythonAstV1Chunker;
+
+impl Chunker for CodePythonAstV1Chunker {
+    fn chunker_version(&self) -> ChunkerVersion {
+        ChunkerVersion(VERSION_LABEL.to_string())
+    }
+
+    fn policy_hash(&self, policy: &ChunkPolicy) -> String {
+        let bytes = serde_json_canonicalizer::to_vec(policy)
+            .expect("canonical JSON serialization of ChunkPolicy must not fail");
+        let hex = blake3::hash(&bytes).to_hex().to_string();
+        hex[..POLICY_HASH_HEX_LEN].to_string()
+    }
+
+    fn chunk(
+        &self,
+        doc: &CanonicalDocument,
+        policy: &ChunkPolicy,
+    ) -> anyhow::Result<Vec<Chunk>> {
+        for b in &doc.blocks {
+            let c = match b {
+                Block::Code(c) => c,
+                _ => anyhow::bail!(
+                    "CodePythonAstV1Chunker only handles code docs (got non-Code block)"
+                ),
+            };
+            if !matches!(c.common.source_span, SourceSpan::Code { .. }) {
+                anyhow::bail!(
+                    "CodePythonAstV1Chunker only handles code docs (got non-Code source_span)"
+                );
+            }
+        }
+
+        let base_policy_hash = self.policy_hash(policy);
+        let chunker_version = self.chunker_version();
+        let mut out: Vec<Chunk> = Vec::new();
+
+        for b in &doc.blocks {
+            let cb = match b {
+                Block::Code(c) => c,
+                _ => unreachable!("validated above"),
+            };
+            let (ls, le, symbol, lang) = match &cb.common.source_span {
+                SourceSpan::Code { line_start, line_end, symbol, lang } => {
+                    (*line_start, *line_end, symbol.clone(), lang.clone())
+                }
+                _ => unreachable!("validated above"),
+            };
+            let block_ids: Vec<BlockId> = vec![cb.common.block_id.clone()];
+            let span_lines = le.saturating_sub(ls) + 1;
+
+            if span_lines <= AST_CHUNK_MAX_LINES {
+                let span = SourceSpan::Code {
+                    line_start: ls,
+                    line_end: le,
+                    symbol: symbol.clone(),
+                    lang: lang.clone(),
+                };
+                out.push(make_chunk(
+                    doc, &chunker_version, &block_ids, &base_policy_hash,
+                    None, span, cb.code.clone(),
+                ));
+            } else {
+                let parts = split_oversize(&cb.code);
+                let n = parts.len();
+                for (i, (off_start, off_end, text)) in parts.into_iter().enumerate() {
+                    let part_ls = ls + off_start;
+                    let part_le = ls + off_end;
+                    let part_sym = symbol
+                        .as_ref()
+                        .map(|s| format!("{s} [part {}/{n}]", i + 1));
+                    let span = SourceSpan::Code {
+                        line_start: part_ls,
+                        line_end: part_le,
+                        symbol: part_sym,
+                        lang: lang.clone(),
+                    };
+                    out.push(make_chunk(
+                        doc, &chunker_version, &block_ids, &base_policy_hash,
+                        Some(part_ls), span, text,
+                    ));
+                }
+            }
+        }
+
+        tracing::debug!(
+            target: "kebab-chunk",
+            doc_id = %doc.doc_id,
+            chunks = out.len(),
+            "code-python-ast-v1 chunked",
+        );
+        Ok(out)
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn make_chunk(
+    doc: &CanonicalDocument,
+    chunker_version: &ChunkerVersion,
+    block_ids: &[BlockId],
+    base_policy_hash: &str,
+    split_key: Option<u32>,
+    span: SourceSpan,
+    text: String,
+) -> Chunk {
+    let id_hash = match split_key {
+        Some(k) => format!("{base_policy_hash}#L{k}"),
+        None => base_policy_hash.to_string(),
+    };
+    let chunk_id = id_for_chunk(&doc.doc_id, chunker_version, block_ids, &id_hash);
+    let token_estimate = text.len().div_ceil(BYTES_PER_TOKEN);
+    Chunk {
+        chunk_id,
+        doc_id: DocumentId(doc.doc_id.0.clone()),
+        block_ids: block_ids.to_vec(),
+        text,
+        heading_path: Vec::new(),
+        source_spans: vec![span],
+        token_estimate,
+        chunker_version: chunker_version.clone(),
+        policy_hash: base_policy_hash.to_string(),
+    }
+}
+
+/// Split an oversize unit at blank-line paragraph boundaries, greedily
+/// gluing paragraphs until ~`AST_CHUNK_MAX_LINES` lines accumulate.
+/// Returns `(line_offset_start, line_offset_end, text)` where offsets are
+/// 0-based within the unit (caller adds the unit's absolute `line_start`).
+fn split_oversize(code: &str) -> Vec<(u32, u32, String)> {
+    let lines: Vec<&str> = code.split('\n').collect();
+    let total = lines.len() as u32;
+    let mut out: Vec<(u32, u32, String)> = Vec::new();
+    let mut start: u32 = 0;
+    while start < total {
+        let mut end = (start + AST_CHUNK_MAX_LINES).min(total);
+        let floor = start + (AST_CHUNK_MAX_LINES * 4 / 5);
+        if end < total {
+            if let Some(b) = (floor.min(end)..end)
+                .rev()
+                .find(|&i| lines[i as usize].trim().is_empty())
+            {
+                end = b + 1;
+            }
+        }
+        let text = lines[start as usize..end as usize].join("\n");
+        out.push((start, end.saturating_sub(1), text));
+        start = end;
+    }
+    if out.is_empty() {
+        out.push((0, total.saturating_sub(1), code.to_string()));
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{
+        Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+        SourceSpan, id_for_block, id_for_doc, AssetId, Lang, Metadata, ParserVersion, Provenance,
+        SourceType, TrustLevel, WorkspacePath,
+    };
+    use time::OffsetDateTime;
+
+    fn code_doc(units: &[(&str, u32, u32, &str)]) -> CanonicalDocument {
+        let wp = WorkspacePath("crates/x/src/a.py".into());
+        let aid = AssetId("a".repeat(64));
+        let pv = ParserVersion("code-python-v1".into());
+        let doc_id = id_for_doc(&wp, &aid, &pv);
+        let blocks = units
+            .iter()
+            .enumerate()
+            .map(|(i, (sym, ls, le, code))| {
+                let span = SourceSpan::Code {
+                    line_start: *ls,
+                    line_end: *le,
+                    symbol: Some((*sym).to_string()),
+                    lang: Some("python".into()),
+                };
+                let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+                Block::Code(CodeBlock {
+                    common: CommonBlock { block_id: bid, heading_path: vec![], source_span: span },
+                    lang: Some("python".into()),
+                    code: (*code).to_string(),
+                })
+            })
+            .collect();
+        CanonicalDocument {
+            doc_id, source_asset_id: aid, workspace_path: wp, title: "a".into(),
+            lang: Lang("und".into()), blocks,
+            metadata: Metadata {
+                aliases: vec![], tags: vec![],
+                created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                source_type: SourceType::Note, trust_level: TrustLevel::Primary,
+                user_id_alias: None, user: Default::default(),
+                repo: Some("kebab".into()), git_branch: Some("main".into()),
+                git_commit: Some("0".repeat(40)), code_lang: Some("python".into()),
+            },
+            provenance: Provenance { events: vec![] },
+            parser_version: pv, schema_version: 1, doc_version: 1,
+            last_chunker_version: None, last_embedding_version: None,
+        }
+    }
+    fn policy() -> ChunkPolicy {
+        ChunkPolicy { target_tokens: 500, overlap_tokens: 80,
+            respect_markdown_headings: false,
+            chunker_version: ChunkerVersion(VERSION_LABEL.into()) }
+    }
+
+    #[test]
+    fn chunker_version_is_code_python_ast_v1() {
+        assert_eq!(CodePythonAstV1Chunker.chunker_version(),
+            ChunkerVersion("code-python-ast-v1".into()));
+    }
+
+    #[test]
+    fn one_chunk_per_unit_preserves_code_span() {
+        let doc = code_doc(&[
+            ("parse", 1, 3, "def parse():\n    pass\n    # x"),
+            ("Foo.double", 5, 7, "def double():\n    #\n    pass"),
+        ]);
+        let chunks = CodePythonAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert_eq!(chunks.len(), 2);
+        for c in &chunks {
+            assert_eq!(c.source_spans.len(), 1);
+            assert!(matches!(c.source_spans[0], SourceSpan::Code { .. }));
+            assert_eq!(c.heading_path, Vec::<String>::new());
+            assert_eq!(c.chunker_version.0, "code-python-ast-v1");
+        }
+        match &chunks[0].source_spans[0] {
+            SourceSpan::Code { symbol, line_start, line_end, .. } => {
+                assert_eq!(symbol.as_deref(), Some("parse"));
+                assert_eq!((*line_start, *line_end), (1, 3));
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    #[test]
+    fn oversize_unit_splits_into_parts_with_unique_ids() {
+        let body = (0..500).map(|i| format!("    x{i} = {i}")).collect::<Vec<_>>().join("\n");
+        let code = format!("def big():\n{body}\n");
+        let doc = code_doc(&[("big", 1, 502, &code)]);
+        let chunks = CodePythonAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert!(chunks.len() >= 2, "oversize unit must split, got {}", chunks.len());
+        for c in &chunks {
+            match &c.source_spans[0] {
+                SourceSpan::Code { symbol, .. } => {
+                    assert!(symbol.as_deref().unwrap().starts_with("big [part "),
+                        "part-numbered symbol, got {symbol:?}");
+                }
+                _ => unreachable!(),
+            }
+        }
+        let mut ids: Vec<&str> = chunks.iter().map(|c| c.chunk_id.0.as_str()).collect();
+        let n = ids.len(); ids.sort(); ids.dedup();
+        assert_eq!(ids.len(), n, "chunk_ids unique across split parts");
+    }
+
+    #[test]
+    fn non_code_doc_errors() {
+        use kebab_core::TextBlock;
+        let mut doc = code_doc(&[("parse", 1, 1, "def parse(): pass")]);
+        doc.blocks = vec![Block::Paragraph(TextBlock {
+            common: CommonBlock {
+                block_id: kebab_core::BlockId("b".into()),
+                heading_path: vec![],
+                source_span: SourceSpan::Line { start: 1, end: 1 },
+            },
+            text: "x".into(), inlines: vec![],
+        })];
+        let err = CodePythonAstV1Chunker.chunk(&doc, &policy()).unwrap_err();
+        assert!(err.to_string().contains("CodePythonAstV1Chunker"));
+    }
+
+    #[test]
+    fn deterministic_chunk_ids_1000() {
+        let doc = code_doc(&[("parse", 1, 2, "def parse(): pass\n")]);
+        let base: Vec<String> = CodePythonAstV1Chunker.chunk(&doc, &policy())
+            .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+        for _ in 0..1000 {
+            let again: Vec<String> = CodePythonAstV1Chunker.chunk(&doc, &policy())
+                .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+            assert_eq!(again, base);
+        }
+    }
+
+    #[test]
+    fn policy_hash_matches_md_heading_v1() {
+        let p = policy();
+        assert_eq!(CodePythonAstV1Chunker.policy_hash(&p),
+            crate::MdHeadingV1Chunker.policy_hash(&p));
+    }
+}

crates/kebab-chunk/src/code_ts_ast_v1.rs

@@ -0,0 +1,322 @@
+//! `code-ts-ast-v1` — maps a tree-sitter-derived TypeScript AST
+//! `CanonicalDocument` (one `Block::Code` per semantic unit, each with
+//! `SourceSpan::Code`) to chunks 1:1. A unit longer than
+//! `AST_CHUNK_MAX_LINES` is split into `<symbol> [part i/N]` sub-chunks
+//! at blank-line paragraph boundaries (design §9.1 oversize fallback).
+//!
+//! tree-sitter is intentionally NOT a dependency here: AST work is
+//! parser-side (`kebab-parse-code`, design §6.3). This chunker only
+//! consumes the `CanonicalDocument`.
+//!
+//! `AST_CHUNK_MAX_LINES` is a constant matching
+//! `IngestCodeCfg::default().ast_chunk_max_lines` (200). Per-medium
+//! config threading needs a chunker registry (P+); same deviation
+//! pattern as `pdf-page-v1`'s pinned `chunker_version`
+//! (`tasks/HOTFIXES.md`).
+
+use kebab_core::{
+    Block, BlockId, CanonicalDocument, Chunk, ChunkPolicy, Chunker, ChunkerVersion, DocumentId,
+    SourceSpan, id_for_chunk,
+};
+
+const VERSION_LABEL: &str = "code-ts-ast-v1";
+const BYTES_PER_TOKEN: usize = 3;
+const POLICY_HASH_HEX_LEN: usize = 16;
+const AST_CHUNK_MAX_LINES: u32 = 200;
+
+#[derive(Clone, Copy, Debug, Default)]
+pub struct CodeTsAstV1Chunker;
+
+impl Chunker for CodeTsAstV1Chunker {
+    fn chunker_version(&self) -> ChunkerVersion {
+        ChunkerVersion(VERSION_LABEL.to_string())
+    }
+
+    fn policy_hash(&self, policy: &ChunkPolicy) -> String {
+        let bytes = serde_json_canonicalizer::to_vec(policy)
+            .expect("canonical JSON serialization of ChunkPolicy must not fail");
+        let hex = blake3::hash(&bytes).to_hex().to_string();
+        hex[..POLICY_HASH_HEX_LEN].to_string()
+    }
+
+    fn chunk(
+        &self,
+        doc: &CanonicalDocument,
+        policy: &ChunkPolicy,
+    ) -> anyhow::Result<Vec<Chunk>> {
+        for b in &doc.blocks {
+            let c = match b {
+                Block::Code(c) => c,
+                _ => anyhow::bail!(
+                    "CodeTsAstV1Chunker only handles code docs (got non-Code block)"
+                ),
+            };
+            if !matches!(c.common.source_span, SourceSpan::Code { .. }) {
+                anyhow::bail!(
+                    "CodeTsAstV1Chunker only handles code docs (got non-Code source_span)"
+                );
+            }
+        }
+
+        let base_policy_hash = self.policy_hash(policy);
+        let chunker_version = self.chunker_version();
+        let mut out: Vec<Chunk> = Vec::new();
+
+        for b in &doc.blocks {
+            let cb = match b {
+                Block::Code(c) => c,
+                _ => unreachable!("validated above"),
+            };
+            let (ls, le, symbol, lang) = match &cb.common.source_span {
+                SourceSpan::Code { line_start, line_end, symbol, lang } => {
+                    (*line_start, *line_end, symbol.clone(), lang.clone())
+                }
+                _ => unreachable!("validated above"),
+            };
+            let block_ids: Vec<BlockId> = vec![cb.common.block_id.clone()];
+            let span_lines = le.saturating_sub(ls) + 1;
+
+            if span_lines <= AST_CHUNK_MAX_LINES {
+                let span = SourceSpan::Code {
+                    line_start: ls,
+                    line_end: le,
+                    symbol: symbol.clone(),
+                    lang: lang.clone(),
+                };
+                out.push(make_chunk(
+                    doc, &chunker_version, &block_ids, &base_policy_hash,
+                    None, span, cb.code.clone(),
+                ));
+            } else {
+                let parts = split_oversize(&cb.code);
+                let n = parts.len();
+                for (i, (off_start, off_end, text)) in parts.into_iter().enumerate() {
+                    let part_ls = ls + off_start;
+                    let part_le = ls + off_end;
+                    let part_sym = symbol
+                        .as_ref()
+                        .map(|s| format!("{s} [part {}/{n}]", i + 1));
+                    let span = SourceSpan::Code {
+                        line_start: part_ls,
+                        line_end: part_le,
+                        symbol: part_sym,
+                        lang: lang.clone(),
+                    };
+                    out.push(make_chunk(
+                        doc, &chunker_version, &block_ids, &base_policy_hash,
+                        Some(part_ls), span, text,
+                    ));
+                }
+            }
+        }
+
+        tracing::debug!(
+            target: "kebab-chunk",
+            doc_id = %doc.doc_id,
+            chunks = out.len(),
+            "code-ts-ast-v1 chunked",
+        );
+        Ok(out)
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn make_chunk(
+    doc: &CanonicalDocument,
+    chunker_version: &ChunkerVersion,
+    block_ids: &[BlockId],
+    base_policy_hash: &str,
+    split_key: Option<u32>,
+    span: SourceSpan,
+    text: String,
+) -> Chunk {
+    let id_hash = match split_key {
+        Some(k) => format!("{base_policy_hash}#L{k}"),
+        None => base_policy_hash.to_string(),
+    };
+    let chunk_id = id_for_chunk(&doc.doc_id, chunker_version, block_ids, &id_hash);
+    let token_estimate = text.len().div_ceil(BYTES_PER_TOKEN);
+    Chunk {
+        chunk_id,
+        doc_id: DocumentId(doc.doc_id.0.clone()),
+        block_ids: block_ids.to_vec(),
+        text,
+        heading_path: Vec::new(),
+        source_spans: vec![span],
+        token_estimate,
+        chunker_version: chunker_version.clone(),
+        policy_hash: base_policy_hash.to_string(),
+    }
+}
+
+/// Split an oversize unit at blank-line paragraph boundaries, greedily
+/// gluing paragraphs until ~`AST_CHUNK_MAX_LINES` lines accumulate.
+/// Returns `(line_offset_start, line_offset_end, text)` where offsets are
+/// 0-based within the unit (caller adds the unit's absolute `line_start`).
+fn split_oversize(code: &str) -> Vec<(u32, u32, String)> {
+    let lines: Vec<&str> = code.split('\n').collect();
+    let total = lines.len() as u32;
+    let mut out: Vec<(u32, u32, String)> = Vec::new();
+    let mut start: u32 = 0;
+    while start < total {
+        let mut end = (start + AST_CHUNK_MAX_LINES).min(total);
+        let floor = start + (AST_CHUNK_MAX_LINES * 4 / 5);
+        if end < total {
+            if let Some(b) = (floor.min(end)..end)
+                .rev()
+                .find(|&i| lines[i as usize].trim().is_empty())
+            {
+                end = b + 1;
+            }
+        }
+        let text = lines[start as usize..end as usize].join("\n");
+        out.push((start, end.saturating_sub(1), text));
+        start = end;
+    }
+    if out.is_empty() {
+        out.push((0, total.saturating_sub(1), code.to_string()));
+    }
+    out
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{
+        Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+        SourceSpan, id_for_block, id_for_doc, AssetId, Lang, Metadata, ParserVersion, Provenance,
+        SourceType, TrustLevel, WorkspacePath,
+    };
+    use time::OffsetDateTime;
+
+    fn code_doc(units: &[(&str, u32, u32, &str)]) -> CanonicalDocument {
+        let wp = WorkspacePath("crates/x/src/a.ts".into());
+        let aid = AssetId("a".repeat(64));
+        let pv = ParserVersion("code-ts-v1".into());
+        let doc_id = id_for_doc(&wp, &aid, &pv);
+        let blocks = units
+            .iter()
+            .enumerate()
+            .map(|(i, (sym, ls, le, code))| {
+                let span = SourceSpan::Code {
+                    line_start: *ls,
+                    line_end: *le,
+                    symbol: Some((*sym).to_string()),
+                    lang: Some("typescript".into()),
+                };
+                let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+                Block::Code(CodeBlock {
+                    common: CommonBlock { block_id: bid, heading_path: vec![], source_span: span },
+                    lang: Some("typescript".into()),
+                    code: (*code).to_string(),
+                })
+            })
+            .collect();
+        CanonicalDocument {
+            doc_id, source_asset_id: aid, workspace_path: wp, title: "a".into(),
+            lang: Lang("und".into()), blocks,
+            metadata: Metadata {
+                aliases: vec![], tags: vec![],
+                created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+                source_type: SourceType::Note, trust_level: TrustLevel::Primary,
+                user_id_alias: None, user: Default::default(),
+                repo: Some("kebab".into()), git_branch: Some("main".into()),
+                git_commit: Some("0".repeat(40)), code_lang: Some("typescript".into()),
+            },
+            provenance: Provenance { events: vec![] },
+            parser_version: pv, schema_version: 1, doc_version: 1,
+            last_chunker_version: None, last_embedding_version: None,
+        }
+    }
+    fn policy() -> ChunkPolicy {
+        ChunkPolicy { target_tokens: 500, overlap_tokens: 80,
+            respect_markdown_headings: false,
+            chunker_version: ChunkerVersion(VERSION_LABEL.into()) }
+    }
+
+    #[test]
+    fn chunker_version_is_code_ts_ast_v1() {
+        assert_eq!(CodeTsAstV1Chunker.chunker_version(),
+            ChunkerVersion("code-ts-ast-v1".into()));
+    }
+
+    #[test]
+    fn one_chunk_per_unit_preserves_code_span() {
+        let doc = code_doc(&[
+            ("parse", 1, 3, "function parse(): void {\n    // x\n}"),
+            ("Foo.double", 5, 7, "function double(): number {\n    //\n    return 0;\n}"),
+        ]);
+        let chunks = CodeTsAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert_eq!(chunks.len(), 2);
+        for c in &chunks {
+            assert_eq!(c.source_spans.len(), 1);
+            assert!(matches!(c.source_spans[0], SourceSpan::Code { .. }));
+            assert_eq!(c.heading_path, Vec::<String>::new());
+            assert_eq!(c.chunker_version.0, "code-ts-ast-v1");
+        }
+        match &chunks[0].source_spans[0] {
+            SourceSpan::Code { symbol, line_start, line_end, .. } => {
+                assert_eq!(symbol.as_deref(), Some("parse"));
+                assert_eq!((*line_start, *line_end), (1, 3));
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    #[test]
+    fn oversize_unit_splits_into_parts_with_unique_ids() {
+        let body = (0..500).map(|i| format!("    const x{i} = {i};")).collect::<Vec<_>>().join("\n");
+        let code = format!("function big(): void {{\n{body}\n}}");
+        let doc = code_doc(&[("big", 1, 502, &code)]);
+        let chunks = CodeTsAstV1Chunker.chunk(&doc, &policy()).unwrap();
+        assert!(chunks.len() >= 2, "oversize unit must split, got {}", chunks.len());
+        for c in &chunks {
+            match &c.source_spans[0] {
+                SourceSpan::Code { symbol, .. } => {
+                    assert!(symbol.as_deref().unwrap().starts_with("big [part "),
+                        "part-numbered symbol, got {symbol:?}");
+                }
+                _ => unreachable!(),
+            }
+        }
+        let mut ids: Vec<&str> = chunks.iter().map(|c| c.chunk_id.0.as_str()).collect();
+        let n = ids.len(); ids.sort(); ids.dedup();
+        assert_eq!(ids.len(), n, "chunk_ids unique across split parts");
+    }
+
+    #[test]
+    fn non_code_doc_errors() {
+        use kebab_core::TextBlock;
+        let mut doc = code_doc(&[("parse", 1, 1, "function parse(): void {}")]);
+        doc.blocks = vec![Block::Paragraph(TextBlock {
+            common: CommonBlock {
+                block_id: kebab_core::BlockId("b".into()),
+                heading_path: vec![],
+                source_span: SourceSpan::Line { start: 1, end: 1 },
+            },
+            text: "x".into(), inlines: vec![],
+        })];
+        let err = CodeTsAstV1Chunker.chunk(&doc, &policy()).unwrap_err();
+        assert!(err.to_string().contains("CodeTsAstV1Chunker"));
+    }
+
+    #[test]
+    fn deterministic_chunk_ids_1000() {
+        let doc = code_doc(&[("parse", 1, 2, "function parse(): void {}\n")]);
+        let base: Vec<String> = CodeTsAstV1Chunker.chunk(&doc, &policy())
+            .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+        for _ in 0..1000 {
+            let again: Vec<String> = CodeTsAstV1Chunker.chunk(&doc, &policy())
+                .unwrap().into_iter().map(|c| c.chunk_id.0).collect();
+            assert_eq!(again, base);
+        }
+    }
+
+    #[test]
+    fn policy_hash_matches_md_heading_v1() {
+        let p = policy();
+        assert_eq!(CodeTsAstV1Chunker.policy_hash(&p),
+            crate::MdHeadingV1Chunker.policy_hash(&p));
+    }
+}

crates/kebab-chunk/src/lib.rs

@@ -15,10 +15,22 @@
 //! embedder, the retriever, the LLM, the RAG layer, or the UI layers.
 //! It consumes `CanonicalDocument` purely through `kb-core` types.
 
+mod code_go_ast_v1;
+mod code_java_ast_v1;
+mod code_js_ast_v1;
+mod code_kotlin_ast_v1;
+mod code_python_ast_v1;
 mod code_rust_ast_v1;
+mod code_ts_ast_v1;
 mod md_heading_v1;
 mod pdf_page_v1;
 
+pub use code_go_ast_v1::CodeGoAstV1Chunker;
+pub use code_java_ast_v1::CodeJavaAstV1Chunker;
+pub use code_js_ast_v1::CodeJsAstV1Chunker;
+pub use code_kotlin_ast_v1::CodeKotlinAstV1Chunker;
+pub use code_python_ast_v1::CodePythonAstV1Chunker;
 pub use code_rust_ast_v1::CodeRustAstV1Chunker;
+pub use code_ts_ast_v1::CodeTsAstV1Chunker;
 pub use md_heading_v1::MdHeadingV1Chunker;
 pub use pdf_page_v1::PdfPageV1Chunker;

crates/kebab-chunk/tests/code_go_ast_snapshot.rs

@@ -0,0 +1,221 @@
+//! Snapshot test pinning the `Vec<Chunk>` JSON for a
+//! representative Go code `CanonicalDocument`.
+//!
+//! This is an integration test. `kebab-parse-code` is intentionally NOT
+//! a dev-dep (design §6.3 / §8 boundary: AST extraction is parser-side).
+//! The `CanonicalDocument` is built inline from hand-crafted `Block::Code`
+//! units, which is the same pattern used in `code_rust_ast_v1.rs`'s
+//! internal `code_doc` test helper.
+//!
+//! Set `UPDATE_SNAPSHOTS=1` to re-bake the baseline.
+
+use std::path::PathBuf;
+
+use kebab_chunk::CodeGoAstV1Chunker;
+use kebab_core::{
+    AssetId, Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+    Lang, Metadata, ParserVersion, Provenance, SourceSpan, SourceType, TrustLevel, WorkspacePath,
+    id_for_block, id_for_doc,
+};
+use serde_json::Value;
+use time::OffsetDateTime;
+
+fn fixtures_dir() -> PathBuf {
+    PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("tests")
+        .join("fixtures")
+}
+
+fn fixed_doc() -> CanonicalDocument {
+    let wp = WorkspacePath("kebab_eval/metrics.go".into());
+    let aid = AssetId("b".repeat(64));
+    // Pin parser_version so doc_id / block_ids are reproducible.
+    let pv = ParserVersion("code-go-v1".into());
+    let doc_id = id_for_doc(&wp, &aid, &pv);
+
+    // Build a >200-line function body to force split_oversize.
+    let big_body: String = {
+        let header = "func BigCompute(data []int) int {\n";
+        let body: String = (0..210u32)
+            .map(|i| format!("\tv{i} := 0\n\tif {i} < len(data) {{\n\t\tv{i} = data[{i}]\n\t}}\n"))
+            .collect();
+        let footer = "\treturn len(data)\n}";
+        format!("{header}{body}{footer}")
+    };
+    let big_line_count = big_body.lines().count() as u32;
+    let big_line_end = 48 + big_line_count - 1;
+
+    // Representative units:
+    //  0. import block                    (lines 1–5,   ≤200)
+    //  1. free fn `ComputeMRR`            (lines 7–12,  ≤200)
+    //  2. struct `MetricsCollector`       (lines 14–20, ≤200)
+    //  3. struct `BaseEvaluator`          (lines 22–30, ≤200)
+    //  4. method `Run`                    (lines 32–38, ≤200)
+    //  5. method `Report`                 (lines 40–46, ≤200)
+    //  6. BigCompute (>200 lines)         to force split_oversize
+    let raw_units: Vec<(&str, u32, u32, String)> = vec![
+        (
+            "imports",
+            1,
+            5,
+            "import (\n\t\"fmt\"\n\t\"os\"\n\t\"strings\"\n)".to_string(),
+        ),
+        (
+            "ComputeMRR",
+            7,
+            12,
+            "func ComputeMRR(scores []float64) float64 {\n\tif len(scores) == 0 {\n\t\treturn 0.0\n\t}\n\t_ = fmt.Sprintf(\"%v\", scores)\n\treturn 1.0 / float64(len(scores))\n}".to_string(),
+        ),
+        (
+            "MetricsCollector",
+            14,
+            20,
+            "type MetricsCollector struct {\n\tScores []float64\n\tLabels []string\n\tCounts map[string]int\n\tTotals map[string]float64\n\tTags   []string\n}".to_string(),
+        ),
+        (
+            "BaseEvaluator",
+            22,
+            30,
+            "type BaseEvaluator struct {\n\tName string\n}\n\nfunc (e *BaseEvaluator) Evaluate(data []string) error {\n\t_ = os.Stderr\n\t_ = strings.Join(data, \",\")\n\treturn nil\n}".to_string(),
+        ),
+        (
+            "MetricsCollector.Run",
+            32,
+            38,
+            "func (m *MetricsCollector) Run(inputs []float64) {\n\tfor _, inp := range inputs {\n\t\tm.Scores = append(\n\t\t\tm.Scores,\n\t\t\tinp,\n\t\t)\n\t}\n}".to_string(),
+        ),
+        (
+            "MetricsCollector.Report",
+            40,
+            46,
+            "func (m *MetricsCollector) Report() map[string]interface{} {\n\treturn map[string]interface{}{\n\t\t\"mean\":  0.0,\n\t\t\"count\": len(m.Scores),\n\t\t\"tags\":  m.Tags,\n\t}\n}".to_string(),
+        ),
+        ("BigCompute", 48, big_line_end, big_body),
+    ];
+
+    let blocks: Vec<Block> = raw_units
+        .iter()
+        .enumerate()
+        .map(|(i, (sym, ls, le, code))| {
+            let span = SourceSpan::Code {
+                line_start: *ls,
+                line_end: *le,
+                symbol: Some((*sym).to_string()),
+                lang: Some("go".into()),
+            };
+            let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+            Block::Code(CodeBlock {
+                common: CommonBlock {
+                    block_id: bid,
+                    heading_path: vec![],
+                    source_span: span,
+                },
+                lang: Some("go".into()),
+                code: code.clone(),
+            })
+        })
+        .collect();
+
+    CanonicalDocument {
+        doc_id,
+        source_asset_id: aid,
+        workspace_path: wp,
+        title: "metrics.go".into(),
+        lang: Lang("und".into()),
+        blocks,
+        metadata: Metadata {
+            aliases: vec![],
+            tags: vec![],
+            created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Default::default(),
+            repo: Some("kebab".into()),
+            git_branch: Some("main".into()),
+            git_commit: Some("0".repeat(40)),
+            code_lang: Some("go".into()),
+        },
+        provenance: Provenance { events: vec![] },
+        parser_version: pv,
+        schema_version: 1,
+        doc_version: 1,
+        last_chunker_version: None,
+        last_embedding_version: None,
+    }
+}
+
+fn fixed_policy() -> ChunkPolicy {
+    ChunkPolicy {
+        target_tokens: 500,
+        overlap_tokens: 80,
+        respect_markdown_headings: false,
+        chunker_version: ChunkerVersion("code-go-ast-v1".into()),
+    }
+}
+
+#[test]
+fn code_go_ast_chunks_snapshot() {
+    let doc = fixed_doc();
+    let policy = fixed_policy();
+
+    let chunks = CodeGoAstV1Chunker.chunk(&doc, &policy).expect("chunk");
+    let actual = serde_json::to_value(&chunks).unwrap();
+
+    let dir = fixtures_dir();
+    let baseline_path = dir.join("code-sample.go.chunks.snapshot.json");
+    let baseline_text = match std::fs::read_to_string(&baseline_path) {
+        Ok(s) => s,
+        Err(_) if std::env::var("UPDATE_SNAPSHOTS").is_ok() => {
+            std::fs::create_dir_all(&dir).unwrap();
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            return;
+        }
+        Err(e) => panic!(
+            "missing baseline {}; run with UPDATE_SNAPSHOTS=1 to create: {e}",
+            baseline_path.display()
+        ),
+    };
+    let expected: Value = serde_json::from_str(&baseline_text).expect("baseline parses as json");
+
+    if actual != expected {
+        if std::env::var("UPDATE_SNAPSHOTS").is_ok() {
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            eprintln!("updated baseline {}", baseline_path.display());
+            return;
+        }
+        let pretty = serde_json::to_string_pretty(&actual).unwrap();
+        panic!(
+            "code-go-ast-v1 chunks snapshot drift\n\
+             --- expected ({}) ---\n{baseline_text}\n\
+             --- actual ---\n{pretty}\n\
+             If intentional, re-run with UPDATE_SNAPSHOTS=1.",
+            baseline_path.display()
+        );
+    }
+}
+
+/// Determinism cross-check: re-running the same pipeline yields the same
+/// chunk_ids byte-for-byte.
+#[test]
+fn code_go_ast_chunks_are_deterministic() {
+    let policy = fixed_policy();
+    let baseline: Vec<String> = CodeGoAstV1Chunker
+        .chunk(&fixed_doc(), &policy)
+        .unwrap()
+        .into_iter()
+        .map(|c| c.chunk_id.0)
+        .collect();
+    for _ in 0..5 {
+        let again: Vec<String> = CodeGoAstV1Chunker
+            .chunk(&fixed_doc(), &policy)
+            .unwrap()
+            .into_iter()
+            .map(|c| c.chunk_id.0)
+            .collect();
+        assert_eq!(again, baseline);
+    }
+}

crates/kebab-chunk/tests/code_java_ast_snapshot.rs

@@ -0,0 +1,221 @@
+//! Snapshot test pinning the `Vec<Chunk>` JSON for a
+//! representative Java code `CanonicalDocument`.
+//!
+//! This is an integration test. `kebab-parse-code` is intentionally NOT
+//! a dev-dep (design §6.3 / §8 boundary: AST extraction is parser-side).
+//! The `CanonicalDocument` is built inline from hand-crafted `Block::Code`
+//! units, which is the same pattern used in `code_rust_ast_v1.rs`'s
+//! internal `code_doc` test helper.
+//!
+//! Set `UPDATE_SNAPSHOTS=1` to re-bake the baseline.
+
+use std::path::PathBuf;
+
+use kebab_chunk::CodeJavaAstV1Chunker;
+use kebab_core::{
+    AssetId, Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+    Lang, Metadata, ParserVersion, Provenance, SourceSpan, SourceType, TrustLevel, WorkspacePath,
+    id_for_block, id_for_doc,
+};
+use serde_json::Value;
+use time::OffsetDateTime;
+
+fn fixtures_dir() -> PathBuf {
+    PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("tests")
+        .join("fixtures")
+}
+
+fn fixed_doc() -> CanonicalDocument {
+    let wp = WorkspacePath("src/main/java/com/example/Metrics.java".into());
+    let aid = AssetId("b".repeat(64));
+    // Pin parser_version so doc_id / block_ids are reproducible.
+    let pv = ParserVersion("code-java-v1".into());
+    let doc_id = id_for_doc(&wp, &aid, &pv);
+
+    // Build a >200-line method body to force split_oversize.
+    let big_body: String = {
+        let header = "public class BigCompute {\n    public int compute(int[] data) {\n";
+        let body: String = (0..210u32)
+            .map(|i| format!("        int v{i} = {i} < data.length ? data[{i}] : 0;\n"))
+            .collect();
+        let footer = "        return data.length;\n    }\n}";
+        format!("{header}{body}{footer}")
+    };
+    let big_line_count = big_body.lines().count() as u32;
+    let big_line_end = 48 + big_line_count - 1;
+
+    // Representative units:
+    //  0. import block                      (lines 1–5,   ≤200)
+    //  1. free method `computeMRR`          (lines 7–12,  ≤200)
+    //  2. class `MetricsCollector`          (lines 14–20, ≤200)
+    //  3. class `BaseEvaluator`             (lines 22–30, ≤200)
+    //  4. method `MetricsCollector.run`     (lines 32–38, ≤200)
+    //  5. method `MetricsCollector.report`  (lines 40–46, ≤200)
+    //  6. BigCompute (>200 lines)           to force split_oversize
+    let raw_units: Vec<(&str, u32, u32, String)> = vec![
+        (
+            "imports",
+            1,
+            5,
+            "import java.util.List;\nimport java.util.Map;\nimport java.util.ArrayList;\nimport java.util.HashMap;\nimport java.util.stream.Collectors;".to_string(),
+        ),
+        (
+            "computeMRR",
+            7,
+            12,
+            "public static double computeMRR(List<Double> scores) {\n    if (scores.isEmpty()) {\n        return 0.0;\n    }\n    return 1.0 / scores.size();\n}".to_string(),
+        ),
+        (
+            "MetricsCollector",
+            14,
+            20,
+            "public class MetricsCollector {\n    private List<Double> scores;\n    private List<String> labels;\n    private Map<String, Integer> counts;\n    private Map<String, Double> totals;\n    private List<String> tags;\n}".to_string(),
+        ),
+        (
+            "BaseEvaluator",
+            22,
+            30,
+            "public class BaseEvaluator {\n    private String name;\n\n    public BaseEvaluator(String name) {\n        this.name = name;\n    }\n\n    public void evaluate(List<String> data) throws Exception {\n        String joined = String.join(\",\", data);\n    }\n}".to_string(),
+        ),
+        (
+            "MetricsCollector.run",
+            32,
+            38,
+            "public void run(List<Double> inputs) {\n    for (Double inp : inputs) {\n        scores.add(\n            inp\n        );\n    }\n}".to_string(),
+        ),
+        (
+            "MetricsCollector.report",
+            40,
+            46,
+            "public Map<String, Object> report() {\n    Map<String, Object> result = new HashMap<>();\n    result.put(\"mean\", 0.0);\n    result.put(\"count\", scores.size());\n    result.put(\"tags\", tags);\n    return result;\n}".to_string(),
+        ),
+        ("BigCompute", 48, big_line_end, big_body),
+    ];
+
+    let blocks: Vec<Block> = raw_units
+        .iter()
+        .enumerate()
+        .map(|(i, (sym, ls, le, code))| {
+            let span = SourceSpan::Code {
+                line_start: *ls,
+                line_end: *le,
+                symbol: Some((*sym).to_string()),
+                lang: Some("java".into()),
+            };
+            let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+            Block::Code(CodeBlock {
+                common: CommonBlock {
+                    block_id: bid,
+                    heading_path: vec![],
+                    source_span: span,
+                },
+                lang: Some("java".into()),
+                code: code.clone(),
+            })
+        })
+        .collect();
+
+    CanonicalDocument {
+        doc_id,
+        source_asset_id: aid,
+        workspace_path: wp,
+        title: "Metrics.java".into(),
+        lang: Lang("und".into()),
+        blocks,
+        metadata: Metadata {
+            aliases: vec![],
+            tags: vec![],
+            created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Default::default(),
+            repo: Some("kebab".into()),
+            git_branch: Some("main".into()),
+            git_commit: Some("0".repeat(40)),
+            code_lang: Some("java".into()),
+        },
+        provenance: Provenance { events: vec![] },
+        parser_version: pv,
+        schema_version: 1,
+        doc_version: 1,
+        last_chunker_version: None,
+        last_embedding_version: None,
+    }
+}
+
+fn fixed_policy() -> ChunkPolicy {
+    ChunkPolicy {
+        target_tokens: 500,
+        overlap_tokens: 80,
+        respect_markdown_headings: false,
+        chunker_version: ChunkerVersion("code-java-ast-v1".into()),
+    }
+}
+
+#[test]
+fn code_java_ast_chunks_snapshot() {
+    let doc = fixed_doc();
+    let policy = fixed_policy();
+
+    let chunks = CodeJavaAstV1Chunker.chunk(&doc, &policy).expect("chunk");
+    let actual = serde_json::to_value(&chunks).unwrap();
+
+    let dir = fixtures_dir();
+    let baseline_path = dir.join("code-sample.java.chunks.snapshot.json");
+    let baseline_text = match std::fs::read_to_string(&baseline_path) {
+        Ok(s) => s,
+        Err(_) if std::env::var("UPDATE_SNAPSHOTS").is_ok() => {
+            std::fs::create_dir_all(&dir).unwrap();
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            return;
+        }
+        Err(e) => panic!(
+            "missing baseline {}; run with UPDATE_SNAPSHOTS=1 to create: {e}",
+            baseline_path.display()
+        ),
+    };
+    let expected: Value = serde_json::from_str(&baseline_text).expect("baseline parses as json");
+
+    if actual != expected {
+        if std::env::var("UPDATE_SNAPSHOTS").is_ok() {
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            eprintln!("updated baseline {}", baseline_path.display());
+            return;
+        }
+        let pretty = serde_json::to_string_pretty(&actual).unwrap();
+        panic!(
+            "code-java-ast-v1 chunks snapshot drift\n\
+             --- expected ({}) ---\n{baseline_text}\n\
+             --- actual ---\n{pretty}\n\
+             If intentional, re-run with UPDATE_SNAPSHOTS=1.",
+            baseline_path.display()
+        );
+    }
+}
+
+/// Determinism cross-check: re-running the same pipeline yields the same
+/// chunk_ids byte-for-byte.
+#[test]
+fn code_java_ast_chunks_are_deterministic() {
+    let policy = fixed_policy();
+    let baseline: Vec<String> = CodeJavaAstV1Chunker
+        .chunk(&fixed_doc(), &policy)
+        .unwrap()
+        .into_iter()
+        .map(|c| c.chunk_id.0)
+        .collect();
+    for _ in 0..5 {
+        let again: Vec<String> = CodeJavaAstV1Chunker
+            .chunk(&fixed_doc(), &policy)
+            .unwrap()
+            .into_iter()
+            .map(|c| c.chunk_id.0)
+            .collect();
+        assert_eq!(again, baseline);
+    }
+}

crates/kebab-chunk/tests/code_js_ast_snapshot.rs

@@ -0,0 +1,221 @@
+//! Snapshot test pinning the `Vec<Chunk>` JSON for a
+//! representative JavaScript code `CanonicalDocument`.
+//!
+//! This is an integration test. `kebab-parse-code` is intentionally NOT
+//! a dev-dep (design §6.3 / §8 boundary: AST extraction is parser-side).
+//! The `CanonicalDocument` is built inline from hand-crafted `Block::Code`
+//! units, which is the same pattern used in `code_rust_ast_v1.rs`'s
+//! internal `code_doc` test helper.
+//!
+//! Set `UPDATE_SNAPSHOTS=1` to re-bake the baseline.
+
+use std::path::PathBuf;
+
+use kebab_chunk::CodeJsAstV1Chunker;
+use kebab_core::{
+    AssetId, Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+    Lang, Metadata, ParserVersion, Provenance, SourceSpan, SourceType, TrustLevel, WorkspacePath,
+    id_for_block, id_for_doc,
+};
+use serde_json::Value;
+use time::OffsetDateTime;
+
+fn fixtures_dir() -> PathBuf {
+    PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("tests")
+        .join("fixtures")
+}
+
+fn fixed_doc() -> CanonicalDocument {
+    let wp = WorkspacePath("src/bar.js".into());
+    let aid = AssetId("b".repeat(64));
+    // Pin parser_version so doc_id / block_ids are reproducible.
+    let pv = ParserVersion("code-js-v1".into());
+    let doc_id = id_for_doc(&wp, &aid, &pv);
+
+    // Build a >200-line function body to force split_oversize.
+    let big_body: String = {
+        let header = "function bigTransform(items) {\n";
+        let body: String = (0..210u32)
+            .map(|i| format!("  const v{i} = items[{i}] !== undefined ? items[{i}] : null;\n"))
+            .collect();
+        let footer = "  return items;\n}";
+        format!("{header}{body}{footer}")
+    };
+    let big_line_count = big_body.lines().count() as u32;
+    let big_line_end = 48 + big_line_count - 1;
+
+    // Representative units:
+    //  0. require/import block       (lines 1–5,   ≤200)
+    //  1. free fn `add`              (lines 7–12,  ≤200)
+    //  2. class `EventBus`           (lines 14–20, ≤200)
+    //  3. class `BaseHandler`        (lines 22–30, ≤200)
+    //  4. method `EventBus.emit`     (lines 32–38, ≤200)
+    //  5. method `EventBus.on`       (lines 40–46, ≤200)
+    //  6. bigTransform (>200 lines)  to force split_oversize
+    let raw_units: Vec<(&str, u32, u32, String)> = vec![
+        (
+            "requires",
+            1,
+            5,
+            "const fs = require('fs');\nconst path = require('path');\nconst { EventEmitter } = require('events');\nconst assert = require('assert');\nconst crypto = require('crypto');".to_string(),
+        ),
+        (
+            "add",
+            7,
+            12,
+            "export function add(a, b) {\n  if (typeof a !== 'number') throw new TypeError('a');\n  if (typeof b !== 'number') throw new TypeError('b');\n  const result = a + b;\n  assert(isFinite(result));\n  return result;\n}".to_string(),
+        ),
+        (
+            "EventBus",
+            14,
+            20,
+            "class EventBus {\n  constructor() {\n    this._handlers = new Map();\n    this._history = [];\n    this._maxHistory = 100;\n    this._seq = 0;\n  }\n}".to_string(),
+        ),
+        (
+            "BaseHandler",
+            22,
+            30,
+            "class BaseHandler {\n  handle(event) {\n    throw new Error('not implemented');\n  }\n  batchHandle(events) {\n    const results = [];\n    for (const ev of events) {\n      results.push(this.handle(ev));\n    }\n    return results;\n  }\n}".to_string(),
+        ),
+        (
+            "EventBus.emit",
+            32,
+            38,
+            "class EventBus {\n  emit(name, payload) {\n    const handlers = this._handlers.get(name) ?? [];\n    for (const h of handlers) {\n      h(payload);\n    }\n    return this;\n  }\n}".to_string(),
+        ),
+        (
+            "EventBus.on",
+            40,
+            46,
+            "class EventBus {\n  on(name, handler) {\n    if (!this._handlers.has(name)) {\n      this._handlers.set(name, []);\n    }\n    this._handlers.get(name).push(handler);\n    return this;\n  }\n}".to_string(),
+        ),
+        ("bigTransform", 48, big_line_end, big_body),
+    ];
+
+    let blocks: Vec<Block> = raw_units
+        .iter()
+        .enumerate()
+        .map(|(i, (sym, ls, le, code))| {
+            let span = SourceSpan::Code {
+                line_start: *ls,
+                line_end: *le,
+                symbol: Some((*sym).to_string()),
+                lang: Some("javascript".into()),
+            };
+            let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+            Block::Code(CodeBlock {
+                common: CommonBlock {
+                    block_id: bid,
+                    heading_path: vec![],
+                    source_span: span,
+                },
+                lang: Some("javascript".into()),
+                code: code.clone(),
+            })
+        })
+        .collect();
+
+    CanonicalDocument {
+        doc_id,
+        source_asset_id: aid,
+        workspace_path: wp,
+        title: "bar.js".into(),
+        lang: Lang("und".into()),
+        blocks,
+        metadata: Metadata {
+            aliases: vec![],
+            tags: vec![],
+            created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Default::default(),
+            repo: Some("kebab".into()),
+            git_branch: Some("main".into()),
+            git_commit: Some("0".repeat(40)),
+            code_lang: Some("javascript".into()),
+        },
+        provenance: Provenance { events: vec![] },
+        parser_version: pv,
+        schema_version: 1,
+        doc_version: 1,
+        last_chunker_version: None,
+        last_embedding_version: None,
+    }
+}
+
+fn fixed_policy() -> ChunkPolicy {
+    ChunkPolicy {
+        target_tokens: 500,
+        overlap_tokens: 80,
+        respect_markdown_headings: false,
+        chunker_version: ChunkerVersion("code-js-ast-v1".into()),
+    }
+}
+
+#[test]
+fn code_js_ast_chunks_snapshot() {
+    let doc = fixed_doc();
+    let policy = fixed_policy();
+
+    let chunks = CodeJsAstV1Chunker.chunk(&doc, &policy).expect("chunk");
+    let actual = serde_json::to_value(&chunks).unwrap();
+
+    let dir = fixtures_dir();
+    let baseline_path = dir.join("code-sample.js.chunks.snapshot.json");
+    let baseline_text = match std::fs::read_to_string(&baseline_path) {
+        Ok(s) => s,
+        Err(_) if std::env::var("UPDATE_SNAPSHOTS").is_ok() => {
+            std::fs::create_dir_all(&dir).unwrap();
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            return;
+        }
+        Err(e) => panic!(
+            "missing baseline {}; run with UPDATE_SNAPSHOTS=1 to create: {e}",
+            baseline_path.display()
+        ),
+    };
+    let expected: Value = serde_json::from_str(&baseline_text).expect("baseline parses as json");
+
+    if actual != expected {
+        if std::env::var("UPDATE_SNAPSHOTS").is_ok() {
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            eprintln!("updated baseline {}", baseline_path.display());
+            return;
+        }
+        let pretty = serde_json::to_string_pretty(&actual).unwrap();
+        panic!(
+            "code-js-ast-v1 chunks snapshot drift\n\
+             --- expected ({}) ---\n{baseline_text}\n\
+             --- actual ---\n{pretty}\n\
+             If intentional, re-run with UPDATE_SNAPSHOTS=1.",
+            baseline_path.display()
+        );
+    }
+}
+
+/// Determinism cross-check: re-running the same pipeline yields the same
+/// chunk_ids byte-for-byte.
+#[test]
+fn code_js_ast_chunks_are_deterministic() {
+    let policy = fixed_policy();
+    let baseline: Vec<String> = CodeJsAstV1Chunker
+        .chunk(&fixed_doc(), &policy)
+        .unwrap()
+        .into_iter()
+        .map(|c| c.chunk_id.0)
+        .collect();
+    for _ in 0..5 {
+        let again: Vec<String> = CodeJsAstV1Chunker
+            .chunk(&fixed_doc(), &policy)
+            .unwrap()
+            .into_iter()
+            .map(|c| c.chunk_id.0)
+            .collect();
+        assert_eq!(again, baseline);
+    }
+}

crates/kebab-chunk/tests/code_kotlin_ast_snapshot.rs

@@ -0,0 +1,221 @@
+//! Snapshot test pinning the `Vec<Chunk>` JSON for a
+//! representative Kotlin code `CanonicalDocument`.
+//!
+//! This is an integration test. `kebab-parse-code` is intentionally NOT
+//! a dev-dep (design §6.3 / §8 boundary: AST extraction is parser-side).
+//! The `CanonicalDocument` is built inline from hand-crafted `Block::Code`
+//! units, which is the same pattern used in `code_rust_ast_v1.rs`'s
+//! internal `code_doc` test helper.
+//!
+//! Set `UPDATE_SNAPSHOTS=1` to re-bake the baseline.
+
+use std::path::PathBuf;
+
+use kebab_chunk::CodeKotlinAstV1Chunker;
+use kebab_core::{
+    AssetId, Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+    Lang, Metadata, ParserVersion, Provenance, SourceSpan, SourceType, TrustLevel, WorkspacePath,
+    id_for_block, id_for_doc,
+};
+use serde_json::Value;
+use time::OffsetDateTime;
+
+fn fixtures_dir() -> PathBuf {
+    PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("tests")
+        .join("fixtures")
+}
+
+fn fixed_doc() -> CanonicalDocument {
+    let wp = WorkspacePath("src/main/kotlin/com/example/Metrics.kt".into());
+    let aid = AssetId("b".repeat(64));
+    // Pin parser_version so doc_id / block_ids are reproducible.
+    let pv = ParserVersion("code-kotlin-v1".into());
+    let doc_id = id_for_doc(&wp, &aid, &pv);
+
+    // Build a >200-line function body to force split_oversize.
+    let big_body: String = {
+        let header = "class BigCompute {\n    fun compute(data: IntArray): Int {\n";
+        let body: String = (0..210u32)
+            .map(|i| format!("        val v{i} = if ({i} < data.size) data[{i}] else 0\n"))
+            .collect();
+        let footer = "        return data.size\n    }\n}";
+        format!("{header}{body}{footer}")
+    };
+    let big_line_count = big_body.lines().count() as u32;
+    let big_line_end = 48 + big_line_count - 1;
+
+    // Representative units:
+    //  0. import block                      (lines 1–5,   ≤200)
+    //  1. top-level fn `computeMRR`         (lines 7–12,  ≤200)
+    //  2. data class `MetricsCollector`     (lines 14–20, ≤200)
+    //  3. class `BaseEvaluator`             (lines 22–30, ≤200)
+    //  4. method `MetricsCollector.run`     (lines 32–38, ≤200)
+    //  5. method `MetricsCollector.report`  (lines 40–46, ≤200)
+    //  6. BigCompute (>200 lines)           to force split_oversize
+    let raw_units: Vec<(&str, u32, u32, String)> = vec![
+        (
+            "imports",
+            1,
+            5,
+            "import kotlin.collections.List\nimport kotlin.collections.Map\nimport kotlin.collections.MutableList\nimport kotlin.collections.MutableMap\nimport kotlin.collections.mutableListOf".to_string(),
+        ),
+        (
+            "computeMRR",
+            7,
+            12,
+            "fun computeMRR(scores: List<Double>): Double {\n    if (scores.isEmpty()) {\n        return 0.0\n    }\n    return 1.0 / scores.size\n}".to_string(),
+        ),
+        (
+            "MetricsCollector",
+            14,
+            20,
+            "data class MetricsCollector(\n    val scores: MutableList<Double> = mutableListOf(),\n    val labels: MutableList<String> = mutableListOf(),\n    val counts: MutableMap<String, Int> = mutableMapOf(),\n    val totals: MutableMap<String, Double> = mutableMapOf(),\n    val tags: MutableList<String> = mutableListOf(),\n)".to_string(),
+        ),
+        (
+            "BaseEvaluator",
+            22,
+            30,
+            "open class BaseEvaluator(val name: String) {\n\n    fun evaluate(data: List<String>) {\n        val joined = data.joinToString(\",\")\n        println(joined)\n    }\n\n    open fun describe(): String = name\n}".to_string(),
+        ),
+        (
+            "MetricsCollector.run",
+            32,
+            38,
+            "fun MetricsCollector.run(inputs: List<Double>) {\n    for (inp in inputs) {\n        scores.add(\n            inp\n        )\n    }\n}".to_string(),
+        ),
+        (
+            "MetricsCollector.report",
+            40,
+            46,
+            "fun MetricsCollector.report(): Map<String, Any> {\n    return mapOf(\n        \"mean\" to 0.0,\n        \"count\" to scores.size,\n        \"tags\" to tags,\n    )\n}".to_string(),
+        ),
+        ("BigCompute", 48, big_line_end, big_body),
+    ];
+
+    let blocks: Vec<Block> = raw_units
+        .iter()
+        .enumerate()
+        .map(|(i, (sym, ls, le, code))| {
+            let span = SourceSpan::Code {
+                line_start: *ls,
+                line_end: *le,
+                symbol: Some((*sym).to_string()),
+                lang: Some("kotlin".into()),
+            };
+            let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+            Block::Code(CodeBlock {
+                common: CommonBlock {
+                    block_id: bid,
+                    heading_path: vec![],
+                    source_span: span,
+                },
+                lang: Some("kotlin".into()),
+                code: code.clone(),
+            })
+        })
+        .collect();
+
+    CanonicalDocument {
+        doc_id,
+        source_asset_id: aid,
+        workspace_path: wp,
+        title: "Metrics.kt".into(),
+        lang: Lang("und".into()),
+        blocks,
+        metadata: Metadata {
+            aliases: vec![],
+            tags: vec![],
+            created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Default::default(),
+            repo: Some("kebab".into()),
+            git_branch: Some("main".into()),
+            git_commit: Some("0".repeat(40)),
+            code_lang: Some("kotlin".into()),
+        },
+        provenance: Provenance { events: vec![] },
+        parser_version: pv,
+        schema_version: 1,
+        doc_version: 1,
+        last_chunker_version: None,
+        last_embedding_version: None,
+    }
+}
+
+fn fixed_policy() -> ChunkPolicy {
+    ChunkPolicy {
+        target_tokens: 500,
+        overlap_tokens: 80,
+        respect_markdown_headings: false,
+        chunker_version: ChunkerVersion("code-kotlin-ast-v1".into()),
+    }
+}
+
+#[test]
+fn code_kotlin_ast_chunks_snapshot() {
+    let doc = fixed_doc();
+    let policy = fixed_policy();
+
+    let chunks = CodeKotlinAstV1Chunker.chunk(&doc, &policy).expect("chunk");
+    let actual = serde_json::to_value(&chunks).unwrap();
+
+    let dir = fixtures_dir();
+    let baseline_path = dir.join("code-sample.kt.chunks.snapshot.json");
+    let baseline_text = match std::fs::read_to_string(&baseline_path) {
+        Ok(s) => s,
+        Err(_) if std::env::var("UPDATE_SNAPSHOTS").is_ok() => {
+            std::fs::create_dir_all(&dir).unwrap();
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            return;
+        }
+        Err(e) => panic!(
+            "missing baseline {}; run with UPDATE_SNAPSHOTS=1 to create: {e}",
+            baseline_path.display()
+        ),
+    };
+    let expected: Value = serde_json::from_str(&baseline_text).expect("baseline parses as json");
+
+    if actual != expected {
+        if std::env::var("UPDATE_SNAPSHOTS").is_ok() {
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            eprintln!("updated baseline {}", baseline_path.display());
+            return;
+        }
+        let pretty = serde_json::to_string_pretty(&actual).unwrap();
+        panic!(
+            "code-kotlin-ast-v1 chunks snapshot drift\n\
+             --- expected ({}) ---\n{baseline_text}\n\
+             --- actual ---\n{pretty}\n\
+             If intentional, re-run with UPDATE_SNAPSHOTS=1.",
+            baseline_path.display()
+        );
+    }
+}
+
+/// Determinism cross-check: re-running the same pipeline yields the same
+/// chunk_ids byte-for-byte.
+#[test]
+fn code_kotlin_ast_chunks_are_deterministic() {
+    let policy = fixed_policy();
+    let baseline: Vec<String> = CodeKotlinAstV1Chunker
+        .chunk(&fixed_doc(), &policy)
+        .unwrap()
+        .into_iter()
+        .map(|c| c.chunk_id.0)
+        .collect();
+    for _ in 0..5 {
+        let again: Vec<String> = CodeKotlinAstV1Chunker
+            .chunk(&fixed_doc(), &policy)
+            .unwrap()
+            .into_iter()
+            .map(|c| c.chunk_id.0)
+            .collect();
+        assert_eq!(again, baseline);
+    }
+}

crates/kebab-chunk/tests/code_python_ast_snapshot.rs

@@ -0,0 +1,221 @@
+//! Snapshot test pinning the `Vec<Chunk>` JSON for a
+//! representative Python code `CanonicalDocument`.
+//!
+//! This is an integration test. `kebab-parse-code` is intentionally NOT
+//! a dev-dep (design §6.3 / §8 boundary: AST extraction is parser-side).
+//! The `CanonicalDocument` is built inline from hand-crafted `Block::Code`
+//! units, which is the same pattern used in `code_rust_ast_v1.rs`'s
+//! internal `code_doc` test helper.
+//!
+//! Set `UPDATE_SNAPSHOTS=1` to re-bake the baseline.
+
+use std::path::PathBuf;
+
+use kebab_chunk::CodePythonAstV1Chunker;
+use kebab_core::{
+    AssetId, Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+    Lang, Metadata, ParserVersion, Provenance, SourceSpan, SourceType, TrustLevel, WorkspacePath,
+    id_for_block, id_for_doc,
+};
+use serde_json::Value;
+use time::OffsetDateTime;
+
+fn fixtures_dir() -> PathBuf {
+    PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("tests")
+        .join("fixtures")
+}
+
+fn fixed_doc() -> CanonicalDocument {
+    let wp = WorkspacePath("kebab_eval/metrics.py".into());
+    let aid = AssetId("b".repeat(64));
+    // Pin parser_version so doc_id / block_ids are reproducible.
+    let pv = ParserVersion("code-python-v1".into());
+    let doc_id = id_for_doc(&wp, &aid, &pv);
+
+    // Build a >200-line function body to force split_oversize.
+    let big_body: String = {
+        let header = "def big_compute(data):\n";
+        let body: String = (0..210u32)
+            .map(|i| format!("    v{i} = data[{i}] if {i} < len(data) else 0\n"))
+            .collect();
+        let footer = "    return sum(data)";
+        format!("{header}{body}{footer}")
+    };
+    let big_line_count = big_body.lines().count() as u32;
+    let big_line_end = 48 + big_line_count - 1;
+
+    // Representative units:
+    //  0. import block               (lines 1–5,   ≤200)
+    //  1. free fn `compute_mrr`      (lines 7–12,  ≤200)
+    //  2. class `MetricsCollector`   (lines 14–20, ≤200)
+    //  3. class `BaseEvaluator`      (lines 22–30, ≤200)
+    //  4. method `run`               (lines 32–38, ≤200)
+    //  5. method `report`            (lines 40–46, ≤200)
+    //  6. big_compute (>200 lines)   to force split_oversize
+    let raw_units: Vec<(&str, u32, u32, String)> = vec![
+        (
+            "imports",
+            1,
+            5,
+            "import os\nimport sys\nfrom typing import List\nfrom pathlib import Path\nfrom collections import defaultdict".to_string(),
+        ),
+        (
+            "compute_mrr",
+            7,
+            12,
+            "def compute_mrr(scores):\n    if not scores:\n        return 0.0\n    return sum(\n        1.0 / r for r in scores\n    ) / len(scores)".to_string(),
+        ),
+        (
+            "MetricsCollector",
+            14,
+            20,
+            "class MetricsCollector:\n    def __init__(self):\n        self.scores = []\n        self.labels = []\n        self.counts = defaultdict(int)\n        self.totals = defaultdict(float)\n        self.tags = []".to_string(),
+        ),
+        (
+            "BaseEvaluator",
+            22,
+            30,
+            "class BaseEvaluator:\n    def evaluate(self, data):\n        raise NotImplementedError\n    def batch_evaluate(self, items):\n        results = []\n        for item in items:\n            results.append(self.evaluate(item))\n        return results\n    def name(self):\n        return type(self).__name__".to_string(),
+        ),
+        (
+            "MetricsCollector.run",
+            32,
+            38,
+            "class MetricsCollector:\n    def run(self, inputs):\n        for inp in inputs:\n            score = self._score(inp)\n            self.scores.append(\n                score\n            )".to_string(),
+        ),
+        (
+            "MetricsCollector.report",
+            40,
+            46,
+            "class MetricsCollector:\n    def report(self):\n        return {\n            'mean': sum(self.scores) / max(len(self.scores), 1),\n            'count': len(self.scores),\n            'tags': self.tags,\n        }".to_string(),
+        ),
+        ("big_compute", 48, big_line_end, big_body),
+    ];
+
+    let blocks: Vec<Block> = raw_units
+        .iter()
+        .enumerate()
+        .map(|(i, (sym, ls, le, code))| {
+            let span = SourceSpan::Code {
+                line_start: *ls,
+                line_end: *le,
+                symbol: Some((*sym).to_string()),
+                lang: Some("python".into()),
+            };
+            let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+            Block::Code(CodeBlock {
+                common: CommonBlock {
+                    block_id: bid,
+                    heading_path: vec![],
+                    source_span: span,
+                },
+                lang: Some("python".into()),
+                code: code.clone(),
+            })
+        })
+        .collect();
+
+    CanonicalDocument {
+        doc_id,
+        source_asset_id: aid,
+        workspace_path: wp,
+        title: "metrics.py".into(),
+        lang: Lang("und".into()),
+        blocks,
+        metadata: Metadata {
+            aliases: vec![],
+            tags: vec![],
+            created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Default::default(),
+            repo: Some("kebab".into()),
+            git_branch: Some("main".into()),
+            git_commit: Some("0".repeat(40)),
+            code_lang: Some("python".into()),
+        },
+        provenance: Provenance { events: vec![] },
+        parser_version: pv,
+        schema_version: 1,
+        doc_version: 1,
+        last_chunker_version: None,
+        last_embedding_version: None,
+    }
+}
+
+fn fixed_policy() -> ChunkPolicy {
+    ChunkPolicy {
+        target_tokens: 500,
+        overlap_tokens: 80,
+        respect_markdown_headings: false,
+        chunker_version: ChunkerVersion("code-python-ast-v1".into()),
+    }
+}
+
+#[test]
+fn code_python_ast_chunks_snapshot() {
+    let doc = fixed_doc();
+    let policy = fixed_policy();
+
+    let chunks = CodePythonAstV1Chunker.chunk(&doc, &policy).expect("chunk");
+    let actual = serde_json::to_value(&chunks).unwrap();
+
+    let dir = fixtures_dir();
+    let baseline_path = dir.join("code-sample.py.chunks.snapshot.json");
+    let baseline_text = match std::fs::read_to_string(&baseline_path) {
+        Ok(s) => s,
+        Err(_) if std::env::var("UPDATE_SNAPSHOTS").is_ok() => {
+            std::fs::create_dir_all(&dir).unwrap();
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            return;
+        }
+        Err(e) => panic!(
+            "missing baseline {}; run with UPDATE_SNAPSHOTS=1 to create: {e}",
+            baseline_path.display()
+        ),
+    };
+    let expected: Value = serde_json::from_str(&baseline_text).expect("baseline parses as json");
+
+    if actual != expected {
+        if std::env::var("UPDATE_SNAPSHOTS").is_ok() {
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            eprintln!("updated baseline {}", baseline_path.display());
+            return;
+        }
+        let pretty = serde_json::to_string_pretty(&actual).unwrap();
+        panic!(
+            "code-python-ast-v1 chunks snapshot drift\n\
+             --- expected ({}) ---\n{baseline_text}\n\
+             --- actual ---\n{pretty}\n\
+             If intentional, re-run with UPDATE_SNAPSHOTS=1.",
+            baseline_path.display()
+        );
+    }
+}
+
+/// Determinism cross-check: re-running the same pipeline yields the same
+/// chunk_ids byte-for-byte.
+#[test]
+fn code_python_ast_chunks_are_deterministic() {
+    let policy = fixed_policy();
+    let baseline: Vec<String> = CodePythonAstV1Chunker
+        .chunk(&fixed_doc(), &policy)
+        .unwrap()
+        .into_iter()
+        .map(|c| c.chunk_id.0)
+        .collect();
+    for _ in 0..5 {
+        let again: Vec<String> = CodePythonAstV1Chunker
+            .chunk(&fixed_doc(), &policy)
+            .unwrap()
+            .into_iter()
+            .map(|c| c.chunk_id.0)
+            .collect();
+        assert_eq!(again, baseline);
+    }
+}

crates/kebab-chunk/tests/code_ts_ast_snapshot.rs

@@ -0,0 +1,221 @@
+//! Snapshot test pinning the `Vec<Chunk>` JSON for a
+//! representative TypeScript code `CanonicalDocument`.
+//!
+//! This is an integration test. `kebab-parse-code` is intentionally NOT
+//! a dev-dep (design §6.3 / §8 boundary: AST extraction is parser-side).
+//! The `CanonicalDocument` is built inline from hand-crafted `Block::Code`
+//! units, which is the same pattern used in `code_rust_ast_v1.rs`'s
+//! internal `code_doc` test helper.
+//!
+//! Set `UPDATE_SNAPSHOTS=1` to re-bake the baseline.
+
+use std::path::PathBuf;
+
+use kebab_chunk::CodeTsAstV1Chunker;
+use kebab_core::{
+    AssetId, Block, CanonicalDocument, ChunkPolicy, Chunker, ChunkerVersion, CodeBlock, CommonBlock,
+    Lang, Metadata, ParserVersion, Provenance, SourceSpan, SourceType, TrustLevel, WorkspacePath,
+    id_for_block, id_for_doc,
+};
+use serde_json::Value;
+use time::OffsetDateTime;
+
+fn fixtures_dir() -> PathBuf {
+    PathBuf::from(env!("CARGO_MANIFEST_DIR"))
+        .join("tests")
+        .join("fixtures")
+}
+
+fn fixed_doc() -> CanonicalDocument {
+    let wp = WorkspacePath("src/Foo.ts".into());
+    let aid = AssetId("b".repeat(64));
+    // Pin parser_version so doc_id / block_ids are reproducible.
+    let pv = ParserVersion("code-ts-v1".into());
+    let doc_id = id_for_doc(&wp, &aid, &pv);
+
+    // Build a >200-line method body to force split_oversize.
+    let big_body: String = {
+        let header = "export class BigProcessor {\n  process(items: string[]): string[] {\n";
+        let body: String = (0..210u32)
+            .map(|i| format!("    const v{i} = items[{i}] ?? '';\n"))
+            .collect();
+        let footer = "    return items;\n  }\n}";
+        format!("{header}{body}{footer}")
+    };
+    let big_line_count = big_body.lines().count() as u32;
+    let big_line_end = 48 + big_line_count - 1;
+
+    // Representative units:
+    //  0. import block               (lines 1–5,   ≤200)
+    //  1. free fn `parseInput`       (lines 7–12,  ≤200)
+    //  2. interface `Frobable`       (lines 14–20, ≤200)
+    //  3. class `Foo`                (lines 22–30, ≤200)
+    //  4. method `Foo.double`        (lines 32–38, ≤200)
+    //  5. method `Foo.triple`        (lines 40–46, ≤200)
+    //  6. BigProcessor (>200 lines)  to force split_oversize
+    let raw_units: Vec<(&str, u32, u32, String)> = vec![
+        (
+            "imports",
+            1,
+            5,
+            "import { readFileSync } from 'fs';\nimport { join } from 'path';\nimport type { Config } from './config';\nimport { Logger } from './logger';\nimport { EventEmitter } from 'events';".to_string(),
+        ),
+        (
+            "parseInput",
+            7,
+            12,
+            "export function parseInput(raw: string): number | null {\n  const trimmed = raw.trim();\n  const n = Number(trimmed);\n  if (isNaN(n)) return null;\n  return n;\n}".to_string(),
+        ),
+        (
+            "Frobable",
+            14,
+            20,
+            "export interface Frobable {\n  frob(): string;\n  frobTwice(): string;\n  readonly name: string;\n  readonly tags: string[];\n  count: number;\n  reset(): void;\n}".to_string(),
+        ),
+        (
+            "Foo",
+            22,
+            30,
+            "export class Foo implements Frobable {\n  constructor(\n    public readonly name: string,\n    public value: number,\n    public tags: string[] = [],\n  ) {}\n  frob(): string { return this.name; }\n  frobTwice(): string { return this.name.repeat(2); }\n  reset(): void { this.value = 0; }\n}".to_string(),
+        ),
+        (
+            "Foo.double",
+            32,
+            38,
+            "export class Foo {\n  double(): number {\n    const result = this.value * 2;\n    if (result > Number.MAX_SAFE_INTEGER) {\n      return Number.MAX_SAFE_INTEGER;\n    }\n    return result;\n  }\n}".to_string(),
+        ),
+        (
+            "Foo.triple",
+            40,
+            46,
+            "export class Foo {\n  triple(): number {\n    const result = this.value * 3;\n    if (result > Number.MAX_SAFE_INTEGER) {\n      return Number.MAX_SAFE_INTEGER;\n    }\n    return result;\n  }\n}".to_string(),
+        ),
+        ("BigProcessor", 48, big_line_end, big_body),
+    ];
+
+    let blocks: Vec<Block> = raw_units
+        .iter()
+        .enumerate()
+        .map(|(i, (sym, ls, le, code))| {
+            let span = SourceSpan::Code {
+                line_start: *ls,
+                line_end: *le,
+                symbol: Some((*sym).to_string()),
+                lang: Some("typescript".into()),
+            };
+            let bid = id_for_block(&doc_id, "code", &[], i as u32, &span);
+            Block::Code(CodeBlock {
+                common: CommonBlock {
+                    block_id: bid,
+                    heading_path: vec![],
+                    source_span: span,
+                },
+                lang: Some("typescript".into()),
+                code: code.clone(),
+            })
+        })
+        .collect();
+
+    CanonicalDocument {
+        doc_id,
+        source_asset_id: aid,
+        workspace_path: wp,
+        title: "Foo.ts".into(),
+        lang: Lang("und".into()),
+        blocks,
+        metadata: Metadata {
+            aliases: vec![],
+            tags: vec![],
+            created_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            updated_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Default::default(),
+            repo: Some("kebab".into()),
+            git_branch: Some("main".into()),
+            git_commit: Some("0".repeat(40)),
+            code_lang: Some("typescript".into()),
+        },
+        provenance: Provenance { events: vec![] },
+        parser_version: pv,
+        schema_version: 1,
+        doc_version: 1,
+        last_chunker_version: None,
+        last_embedding_version: None,
+    }
+}
+
+fn fixed_policy() -> ChunkPolicy {
+    ChunkPolicy {
+        target_tokens: 500,
+        overlap_tokens: 80,
+        respect_markdown_headings: false,
+        chunker_version: ChunkerVersion("code-ts-ast-v1".into()),
+    }
+}
+
+#[test]
+fn code_ts_ast_chunks_snapshot() {
+    let doc = fixed_doc();
+    let policy = fixed_policy();
+
+    let chunks = CodeTsAstV1Chunker.chunk(&doc, &policy).expect("chunk");
+    let actual = serde_json::to_value(&chunks).unwrap();
+
+    let dir = fixtures_dir();
+    let baseline_path = dir.join("code-sample.ts.chunks.snapshot.json");
+    let baseline_text = match std::fs::read_to_string(&baseline_path) {
+        Ok(s) => s,
+        Err(_) if std::env::var("UPDATE_SNAPSHOTS").is_ok() => {
+            std::fs::create_dir_all(&dir).unwrap();
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            return;
+        }
+        Err(e) => panic!(
+            "missing baseline {}; run with UPDATE_SNAPSHOTS=1 to create: {e}",
+            baseline_path.display()
+        ),
+    };
+    let expected: Value = serde_json::from_str(&baseline_text).expect("baseline parses as json");
+
+    if actual != expected {
+        if std::env::var("UPDATE_SNAPSHOTS").is_ok() {
+            let pretty = serde_json::to_string_pretty(&actual).unwrap();
+            std::fs::write(&baseline_path, format!("{pretty}\n")).unwrap();
+            eprintln!("updated baseline {}", baseline_path.display());
+            return;
+        }
+        let pretty = serde_json::to_string_pretty(&actual).unwrap();
+        panic!(
+            "code-ts-ast-v1 chunks snapshot drift\n\
+             --- expected ({}) ---\n{baseline_text}\n\
+             --- actual ---\n{pretty}\n\
+             If intentional, re-run with UPDATE_SNAPSHOTS=1.",
+            baseline_path.display()
+        );
+    }
+}
+
+/// Determinism cross-check: re-running the same pipeline yields the same
+/// chunk_ids byte-for-byte.
+#[test]
+fn code_ts_ast_chunks_are_deterministic() {
+    let policy = fixed_policy();
+    let baseline: Vec<String> = CodeTsAstV1Chunker
+        .chunk(&fixed_doc(), &policy)
+        .unwrap()
+        .into_iter()
+        .map(|c| c.chunk_id.0)
+        .collect();
+    for _ in 0..5 {
+        let again: Vec<String> = CodeTsAstV1Chunker
+            .chunk(&fixed_doc(), &policy)
+            .unwrap()
+            .into_iter()
+            .map(|c| c.chunk_id.0)
+            .collect();
+        assert_eq!(again, baseline);
+    }
+}

crates/kebab-chunk/tests/fixtures/code-sample.go.chunks.snapshot.json

@@ -0,0 +1,233 @@
+[
+  {
+    "block_ids": [
+      "c182bf37e32c7fc1b868bd617f8eaf66"
+    ],
+    "chunk_id": "43de518d946dc18ec040ae20d74e0cff",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 5,
+        "line_start": 1,
+        "symbol": "imports"
+      }
+    ],
+    "text": "import (\n\t\"fmt\"\n\t\"os\"\n\t\"strings\"\n)",
+    "token_estimate": 12
+  },
+  {
+    "block_ids": [
+      "c9992cdcfdf3c2a7700a4abc4782a8a4"
+    ],
+    "chunk_id": "af4c382a83f1e8cdea495d8b33c11abc",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 12,
+        "line_start": 7,
+        "symbol": "ComputeMRR"
+      }
+    ],
+    "text": "func ComputeMRR(scores []float64) float64 {\n\tif len(scores) == 0 {\n\t\treturn 0.0\n\t}\n\t_ = fmt.Sprintf(\"%v\", scores)\n\treturn 1.0 / float64(len(scores))\n}",
+    "token_estimate": 50
+  },
+  {
+    "block_ids": [
+      "5f18dc3e79fe946ba05d32c3bfc00684"
+    ],
+    "chunk_id": "4be6d8f180bc19b8651877e5264852ac",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 20,
+        "line_start": 14,
+        "symbol": "MetricsCollector"
+      }
+    ],
+    "text": "type MetricsCollector struct {\n\tScores []float64\n\tLabels []string\n\tCounts map[string]int\n\tTotals map[string]float64\n\tTags   []string\n}",
+    "token_estimate": 45
+  },
+  {
+    "block_ids": [
+      "3009cc022ca832c323393e4f9bcdb388"
+    ],
+    "chunk_id": "3ae182f4c6d304ee7f0aaf447142f948",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 30,
+        "line_start": 22,
+        "symbol": "BaseEvaluator"
+      }
+    ],
+    "text": "type BaseEvaluator struct {\n\tName string\n}\n\nfunc (e *BaseEvaluator) Evaluate(data []string) error {\n\t_ = os.Stderr\n\t_ = strings.Join(data, \",\")\n\treturn nil\n}",
+    "token_estimate": 53
+  },
+  {
+    "block_ids": [
+      "e0e83d1d7f9327a1902ae9a8f67c1f1c"
+    ],
+    "chunk_id": "b962f14980e756bb8ba514e2282756cd",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 38,
+        "line_start": 32,
+        "symbol": "MetricsCollector.Run"
+      }
+    ],
+    "text": "func (m *MetricsCollector) Run(inputs []float64) {\n\tfor _, inp := range inputs {\n\t\tm.Scores = append(\n\t\t\tm.Scores,\n\t\t\tinp,\n\t\t)\n\t}\n}",
+    "token_estimate": 44
+  },
+  {
+    "block_ids": [
+      "0e6a572bc3fe2bd6d173fe614bd1b763"
+    ],
+    "chunk_id": "441c695e990e7f49188068433e313e87",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 46,
+        "line_start": 40,
+        "symbol": "MetricsCollector.Report"
+      }
+    ],
+    "text": "func (m *MetricsCollector) Report() map[string]interface{} {\n\treturn map[string]interface{}{\n\t\t\"mean\":  0.0,\n\t\t\"count\": len(m.Scores),\n\t\t\"tags\":  m.Tags,\n\t}\n}",
+    "token_estimate": 53
+  },
+  {
+    "block_ids": [
+      "5d269745b2e5dbdcbef0c09ba54b0bd6"
+    ],
+    "chunk_id": "7a942d871c588ec69426290561f05179",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 247,
+        "line_start": 48,
+        "symbol": "BigCompute [part 1/5]"
+      }
+    ],
+    "text": "func BigCompute(data []int) int {\n\tv0 := 0\n\tif 0 < len(data) {\n\t\tv0 = data[0]\n\t}\n\tv1 := 0\n\tif 1 < len(data) {\n\t\tv1 = data[1]\n\t}\n\tv2 := 0\n\tif 2 < len(data) {\n\t\tv2 = data[2]\n\t}\n\tv3 := 0\n\tif 3 < len(data) {\n\t\tv3 = data[3]\n\t}\n\tv4 := 0\n\tif 4 < len(data) {\n\t\tv4 = data[4]\n\t}\n\tv5 := 0\n\tif 5 < len(data) {\n\t\tv5 = data[5]\n\t}\n\tv6 := 0\n\tif 6 < len(data) {\n\t\tv6 = data[6]\n\t}\n\tv7 := 0\n\tif 7 < len(data) {\n\t\tv7 = data[7]\n\t}\n\tv8 := 0\n\tif 8 < len(data) {\n\t\tv8 = data[8]\n\t}\n\tv9 := 0\n\tif 9 < len(data) {\n\t\tv9 = data[9]\n\t}\n\tv10 := 0\n\tif 10 < len(data) {\n\t\tv10 = data[10]\n\t}\n\tv11 := 0\n\tif 11 < len(data) {\n\t\tv11 = data[11]\n\t}\n\tv12 := 0\n\tif 12 < len(data) {\n\t\tv12 = data[12]\n\t}\n\tv13 := 0\n\tif 13 < len(data) {\n\t\tv13 = data[13]\n\t}\n\tv14 := 0\n\tif 14 < len(data) {\n\t\tv14 = data[14]\n\t}\n\tv15 := 0\n\tif 15 < len(data) {\n\t\tv15 = data[15]\n\t}\n\tv16 := 0\n\tif 16 < len(data) {\n\t\tv16 = data[16]\n\t}\n\tv17 := 0\n\tif 17 < len(data) {\n\t\tv17 = data[17]\n\t}\n\tv18 := 0\n\tif 18 < len(data) {\n\t\tv18 = data[18]\n\t}\n\tv19 := 0\n\tif 19 < len(data) {\n\t\tv19 = data[19]\n\t}\n\tv20 := 0\n\tif 20 < len(data) {\n\t\tv20 = data[20]\n\t}\n\tv21 := 0\n\tif 21 < len(data) {\n\t\tv21 = data[21]\n\t}\n\tv22 := 0\n\tif 22 < len(data) {\n\t\tv22 = data[22]\n\t}\n\tv23 := 0\n\tif 23 < len(data) {\n\t\tv23 = data[23]\n\t}\n\tv24 := 0\n\tif 24 < len(data) {\n\t\tv24 = data[24]\n\t}\n\tv25 := 0\n\tif 25 < len(data) {\n\t\tv25 = data[25]\n\t}\n\tv26 := 0\n\tif 26 < len(data) {\n\t\tv26 = data[26]\n\t}\n\tv27 := 0\n\tif 27 < len(data) {\n\t\tv27 = data[27]\n\t}\n\tv28 := 0\n\tif 28 < len(data) {\n\t\tv28 = data[28]\n\t}\n\tv29 := 0\n\tif 29 < len(data) {\n\t\tv29 = data[29]\n\t}\n\tv30 := 0\n\tif 30 < len(data) {\n\t\tv30 = data[30]\n\t}\n\tv31 := 0\n\tif 31 < len(data) {\n\t\tv31 = data[31]\n\t}\n\tv32 := 0\n\tif 32 < len(data) {\n\t\tv32 = data[32]\n\t}\n\tv33 := 0\n\tif 33 < len(data) {\n\t\tv33 = data[33]\n\t}\n\tv34 := 0\n\tif 34 < len(data) {\n\t\tv34 = data[34]\n\t}\n\tv35 := 0\n\tif 35 < len(data) {\n\t\tv35 = data[35]\n\t}\n\tv36 := 0\n\tif 36 < len(data) {\n\t\tv36 = data[36]\n\t}\n\tv37 := 0\n\tif 37 < len(data) {\n\t\tv37 = data[37]\n\t}\n\tv38 := 0\n\tif 38 < len(data) {\n\t\tv38 = data[38]\n\t}\n\tv39 := 0\n\tif 39 < len(data) {\n\t\tv39 = data[39]\n\t}\n\tv40 := 0\n\tif 40 < len(data) {\n\t\tv40 = data[40]\n\t}\n\tv41 := 0\n\tif 41 < len(data) {\n\t\tv41 = data[41]\n\t}\n\tv42 := 0\n\tif 42 < len(data) {\n\t\tv42 = data[42]\n\t}\n\tv43 := 0\n\tif 43 < len(data) {\n\t\tv43 = data[43]\n\t}\n\tv44 := 0\n\tif 44 < len(data) {\n\t\tv44 = data[44]\n\t}\n\tv45 := 0\n\tif 45 < len(data) {\n\t\tv45 = data[45]\n\t}\n\tv46 := 0\n\tif 46 < len(data) {\n\t\tv46 = data[46]\n\t}\n\tv47 := 0\n\tif 47 < len(data) {\n\t\tv47 = data[47]\n\t}\n\tv48 := 0\n\tif 48 < len(data) {\n\t\tv48 = data[48]\n\t}\n\tv49 := 0\n\tif 49 < len(data) {\n\t\tv49 = data[49]",
+    "token_estimate": 847
+  },
+  {
+    "block_ids": [
+      "5d269745b2e5dbdcbef0c09ba54b0bd6"
+    ],
+    "chunk_id": "3f44ba43c9415652e2705bb667776e76",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 447,
+        "line_start": 248,
+        "symbol": "BigCompute [part 2/5]"
+      }
+    ],
+    "text": "\t}\n\tv50 := 0\n\tif 50 < len(data) {\n\t\tv50 = data[50]\n\t}\n\tv51 := 0\n\tif 51 < len(data) {\n\t\tv51 = data[51]\n\t}\n\tv52 := 0\n\tif 52 < len(data) {\n\t\tv52 = data[52]\n\t}\n\tv53 := 0\n\tif 53 < len(data) {\n\t\tv53 = data[53]\n\t}\n\tv54 := 0\n\tif 54 < len(data) {\n\t\tv54 = data[54]\n\t}\n\tv55 := 0\n\tif 55 < len(data) {\n\t\tv55 = data[55]\n\t}\n\tv56 := 0\n\tif 56 < len(data) {\n\t\tv56 = data[56]\n\t}\n\tv57 := 0\n\tif 57 < len(data) {\n\t\tv57 = data[57]\n\t}\n\tv58 := 0\n\tif 58 < len(data) {\n\t\tv58 = data[58]\n\t}\n\tv59 := 0\n\tif 59 < len(data) {\n\t\tv59 = data[59]\n\t}\n\tv60 := 0\n\tif 60 < len(data) {\n\t\tv60 = data[60]\n\t}\n\tv61 := 0\n\tif 61 < len(data) {\n\t\tv61 = data[61]\n\t}\n\tv62 := 0\n\tif 62 < len(data) {\n\t\tv62 = data[62]\n\t}\n\tv63 := 0\n\tif 63 < len(data) {\n\t\tv63 = data[63]\n\t}\n\tv64 := 0\n\tif 64 < len(data) {\n\t\tv64 = data[64]\n\t}\n\tv65 := 0\n\tif 65 < len(data) {\n\t\tv65 = data[65]\n\t}\n\tv66 := 0\n\tif 66 < len(data) {\n\t\tv66 = data[66]\n\t}\n\tv67 := 0\n\tif 67 < len(data) {\n\t\tv67 = data[67]\n\t}\n\tv68 := 0\n\tif 68 < len(data) {\n\t\tv68 = data[68]\n\t}\n\tv69 := 0\n\tif 69 < len(data) {\n\t\tv69 = data[69]\n\t}\n\tv70 := 0\n\tif 70 < len(data) {\n\t\tv70 = data[70]\n\t}\n\tv71 := 0\n\tif 71 < len(data) {\n\t\tv71 = data[71]\n\t}\n\tv72 := 0\n\tif 72 < len(data) {\n\t\tv72 = data[72]\n\t}\n\tv73 := 0\n\tif 73 < len(data) {\n\t\tv73 = data[73]\n\t}\n\tv74 := 0\n\tif 74 < len(data) {\n\t\tv74 = data[74]\n\t}\n\tv75 := 0\n\tif 75 < len(data) {\n\t\tv75 = data[75]\n\t}\n\tv76 := 0\n\tif 76 < len(data) {\n\t\tv76 = data[76]\n\t}\n\tv77 := 0\n\tif 77 < len(data) {\n\t\tv77 = data[77]\n\t}\n\tv78 := 0\n\tif 78 < len(data) {\n\t\tv78 = data[78]\n\t}\n\tv79 := 0\n\tif 79 < len(data) {\n\t\tv79 = data[79]\n\t}\n\tv80 := 0\n\tif 80 < len(data) {\n\t\tv80 = data[80]\n\t}\n\tv81 := 0\n\tif 81 < len(data) {\n\t\tv81 = data[81]\n\t}\n\tv82 := 0\n\tif 82 < len(data) {\n\t\tv82 = data[82]\n\t}\n\tv83 := 0\n\tif 83 < len(data) {\n\t\tv83 = data[83]\n\t}\n\tv84 := 0\n\tif 84 < len(data) {\n\t\tv84 = data[84]\n\t}\n\tv85 := 0\n\tif 85 < len(data) {\n\t\tv85 = data[85]\n\t}\n\tv86 := 0\n\tif 86 < len(data) {\n\t\tv86 = data[86]\n\t}\n\tv87 := 0\n\tif 87 < len(data) {\n\t\tv87 = data[87]\n\t}\n\tv88 := 0\n\tif 88 < len(data) {\n\t\tv88 = data[88]\n\t}\n\tv89 := 0\n\tif 89 < len(data) {\n\t\tv89 = data[89]\n\t}\n\tv90 := 0\n\tif 90 < len(data) {\n\t\tv90 = data[90]\n\t}\n\tv91 := 0\n\tif 91 < len(data) {\n\t\tv91 = data[91]\n\t}\n\tv92 := 0\n\tif 92 < len(data) {\n\t\tv92 = data[92]\n\t}\n\tv93 := 0\n\tif 93 < len(data) {\n\t\tv93 = data[93]\n\t}\n\tv94 := 0\n\tif 94 < len(data) {\n\t\tv94 = data[94]\n\t}\n\tv95 := 0\n\tif 95 < len(data) {\n\t\tv95 = data[95]\n\t}\n\tv96 := 0\n\tif 96 < len(data) {\n\t\tv96 = data[96]\n\t}\n\tv97 := 0\n\tif 97 < len(data) {\n\t\tv97 = data[97]\n\t}\n\tv98 := 0\n\tif 98 < len(data) {\n\t\tv98 = data[98]\n\t}\n\tv99 := 0\n\tif 99 < len(data) {\n\t\tv99 = data[99]",
+    "token_estimate": 850
+  },
+  {
+    "block_ids": [
+      "5d269745b2e5dbdcbef0c09ba54b0bd6"
+    ],
+    "chunk_id": "e4763e10f059d97f40c2932761b56c3e",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 647,
+        "line_start": 448,
+        "symbol": "BigCompute [part 3/5]"
+      }
+    ],
+    "text": "\t}\n\tv100 := 0\n\tif 100 < len(data) {\n\t\tv100 = data[100]\n\t}\n\tv101 := 0\n\tif 101 < len(data) {\n\t\tv101 = data[101]\n\t}\n\tv102 := 0\n\tif 102 < len(data) {\n\t\tv102 = data[102]\n\t}\n\tv103 := 0\n\tif 103 < len(data) {\n\t\tv103 = data[103]\n\t}\n\tv104 := 0\n\tif 104 < len(data) {\n\t\tv104 = data[104]\n\t}\n\tv105 := 0\n\tif 105 < len(data) {\n\t\tv105 = data[105]\n\t}\n\tv106 := 0\n\tif 106 < len(data) {\n\t\tv106 = data[106]\n\t}\n\tv107 := 0\n\tif 107 < len(data) {\n\t\tv107 = data[107]\n\t}\n\tv108 := 0\n\tif 108 < len(data) {\n\t\tv108 = data[108]\n\t}\n\tv109 := 0\n\tif 109 < len(data) {\n\t\tv109 = data[109]\n\t}\n\tv110 := 0\n\tif 110 < len(data) {\n\t\tv110 = data[110]\n\t}\n\tv111 := 0\n\tif 111 < len(data) {\n\t\tv111 = data[111]\n\t}\n\tv112 := 0\n\tif 112 < len(data) {\n\t\tv112 = data[112]\n\t}\n\tv113 := 0\n\tif 113 < len(data) {\n\t\tv113 = data[113]\n\t}\n\tv114 := 0\n\tif 114 < len(data) {\n\t\tv114 = data[114]\n\t}\n\tv115 := 0\n\tif 115 < len(data) {\n\t\tv115 = data[115]\n\t}\n\tv116 := 0\n\tif 116 < len(data) {\n\t\tv116 = data[116]\n\t}\n\tv117 := 0\n\tif 117 < len(data) {\n\t\tv117 = data[117]\n\t}\n\tv118 := 0\n\tif 118 < len(data) {\n\t\tv118 = data[118]\n\t}\n\tv119 := 0\n\tif 119 < len(data) {\n\t\tv119 = data[119]\n\t}\n\tv120 := 0\n\tif 120 < len(data) {\n\t\tv120 = data[120]\n\t}\n\tv121 := 0\n\tif 121 < len(data) {\n\t\tv121 = data[121]\n\t}\n\tv122 := 0\n\tif 122 < len(data) {\n\t\tv122 = data[122]\n\t}\n\tv123 := 0\n\tif 123 < len(data) {\n\t\tv123 = data[123]\n\t}\n\tv124 := 0\n\tif 124 < len(data) {\n\t\tv124 = data[124]\n\t}\n\tv125 := 0\n\tif 125 < len(data) {\n\t\tv125 = data[125]\n\t}\n\tv126 := 0\n\tif 126 < len(data) {\n\t\tv126 = data[126]\n\t}\n\tv127 := 0\n\tif 127 < len(data) {\n\t\tv127 = data[127]\n\t}\n\tv128 := 0\n\tif 128 < len(data) {\n\t\tv128 = data[128]\n\t}\n\tv129 := 0\n\tif 129 < len(data) {\n\t\tv129 = data[129]\n\t}\n\tv130 := 0\n\tif 130 < len(data) {\n\t\tv130 = data[130]\n\t}\n\tv131 := 0\n\tif 131 < len(data) {\n\t\tv131 = data[131]\n\t}\n\tv132 := 0\n\tif 132 < len(data) {\n\t\tv132 = data[132]\n\t}\n\tv133 := 0\n\tif 133 < len(data) {\n\t\tv133 = data[133]\n\t}\n\tv134 := 0\n\tif 134 < len(data) {\n\t\tv134 = data[134]\n\t}\n\tv135 := 0\n\tif 135 < len(data) {\n\t\tv135 = data[135]\n\t}\n\tv136 := 0\n\tif 136 < len(data) {\n\t\tv136 = data[136]\n\t}\n\tv137 := 0\n\tif 137 < len(data) {\n\t\tv137 = data[137]\n\t}\n\tv138 := 0\n\tif 138 < len(data) {\n\t\tv138 = data[138]\n\t}\n\tv139 := 0\n\tif 139 < len(data) {\n\t\tv139 = data[139]\n\t}\n\tv140 := 0\n\tif 140 < len(data) {\n\t\tv140 = data[140]\n\t}\n\tv141 := 0\n\tif 141 < len(data) {\n\t\tv141 = data[141]\n\t}\n\tv142 := 0\n\tif 142 < len(data) {\n\t\tv142 = data[142]\n\t}\n\tv143 := 0\n\tif 143 < len(data) {\n\t\tv143 = data[143]\n\t}\n\tv144 := 0\n\tif 144 < len(data) {\n\t\tv144 = data[144]\n\t}\n\tv145 := 0\n\tif 145 < len(data) {\n\t\tv145 = data[145]\n\t}\n\tv146 := 0\n\tif 146 < len(data) {\n\t\tv146 = data[146]\n\t}\n\tv147 := 0\n\tif 147 < len(data) {\n\t\tv147 = data[147]\n\t}\n\tv148 := 0\n\tif 148 < len(data) {\n\t\tv148 = data[148]\n\t}\n\tv149 := 0\n\tif 149 < len(data) {\n\t\tv149 = data[149]",
+    "token_estimate": 917
+  },
+  {
+    "block_ids": [
+      "5d269745b2e5dbdcbef0c09ba54b0bd6"
+    ],
+    "chunk_id": "24176c911d0bacf9a29fa7f8251f5036",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 847,
+        "line_start": 648,
+        "symbol": "BigCompute [part 4/5]"
+      }
+    ],
+    "text": "\t}\n\tv150 := 0\n\tif 150 < len(data) {\n\t\tv150 = data[150]\n\t}\n\tv151 := 0\n\tif 151 < len(data) {\n\t\tv151 = data[151]\n\t}\n\tv152 := 0\n\tif 152 < len(data) {\n\t\tv152 = data[152]\n\t}\n\tv153 := 0\n\tif 153 < len(data) {\n\t\tv153 = data[153]\n\t}\n\tv154 := 0\n\tif 154 < len(data) {\n\t\tv154 = data[154]\n\t}\n\tv155 := 0\n\tif 155 < len(data) {\n\t\tv155 = data[155]\n\t}\n\tv156 := 0\n\tif 156 < len(data) {\n\t\tv156 = data[156]\n\t}\n\tv157 := 0\n\tif 157 < len(data) {\n\t\tv157 = data[157]\n\t}\n\tv158 := 0\n\tif 158 < len(data) {\n\t\tv158 = data[158]\n\t}\n\tv159 := 0\n\tif 159 < len(data) {\n\t\tv159 = data[159]\n\t}\n\tv160 := 0\n\tif 160 < len(data) {\n\t\tv160 = data[160]\n\t}\n\tv161 := 0\n\tif 161 < len(data) {\n\t\tv161 = data[161]\n\t}\n\tv162 := 0\n\tif 162 < len(data) {\n\t\tv162 = data[162]\n\t}\n\tv163 := 0\n\tif 163 < len(data) {\n\t\tv163 = data[163]\n\t}\n\tv164 := 0\n\tif 164 < len(data) {\n\t\tv164 = data[164]\n\t}\n\tv165 := 0\n\tif 165 < len(data) {\n\t\tv165 = data[165]\n\t}\n\tv166 := 0\n\tif 166 < len(data) {\n\t\tv166 = data[166]\n\t}\n\tv167 := 0\n\tif 167 < len(data) {\n\t\tv167 = data[167]\n\t}\n\tv168 := 0\n\tif 168 < len(data) {\n\t\tv168 = data[168]\n\t}\n\tv169 := 0\n\tif 169 < len(data) {\n\t\tv169 = data[169]\n\t}\n\tv170 := 0\n\tif 170 < len(data) {\n\t\tv170 = data[170]\n\t}\n\tv171 := 0\n\tif 171 < len(data) {\n\t\tv171 = data[171]\n\t}\n\tv172 := 0\n\tif 172 < len(data) {\n\t\tv172 = data[172]\n\t}\n\tv173 := 0\n\tif 173 < len(data) {\n\t\tv173 = data[173]\n\t}\n\tv174 := 0\n\tif 174 < len(data) {\n\t\tv174 = data[174]\n\t}\n\tv175 := 0\n\tif 175 < len(data) {\n\t\tv175 = data[175]\n\t}\n\tv176 := 0\n\tif 176 < len(data) {\n\t\tv176 = data[176]\n\t}\n\tv177 := 0\n\tif 177 < len(data) {\n\t\tv177 = data[177]\n\t}\n\tv178 := 0\n\tif 178 < len(data) {\n\t\tv178 = data[178]\n\t}\n\tv179 := 0\n\tif 179 < len(data) {\n\t\tv179 = data[179]\n\t}\n\tv180 := 0\n\tif 180 < len(data) {\n\t\tv180 = data[180]\n\t}\n\tv181 := 0\n\tif 181 < len(data) {\n\t\tv181 = data[181]\n\t}\n\tv182 := 0\n\tif 182 < len(data) {\n\t\tv182 = data[182]\n\t}\n\tv183 := 0\n\tif 183 < len(data) {\n\t\tv183 = data[183]\n\t}\n\tv184 := 0\n\tif 184 < len(data) {\n\t\tv184 = data[184]\n\t}\n\tv185 := 0\n\tif 185 < len(data) {\n\t\tv185 = data[185]\n\t}\n\tv186 := 0\n\tif 186 < len(data) {\n\t\tv186 = data[186]\n\t}\n\tv187 := 0\n\tif 187 < len(data) {\n\t\tv187 = data[187]\n\t}\n\tv188 := 0\n\tif 188 < len(data) {\n\t\tv188 = data[188]\n\t}\n\tv189 := 0\n\tif 189 < len(data) {\n\t\tv189 = data[189]\n\t}\n\tv190 := 0\n\tif 190 < len(data) {\n\t\tv190 = data[190]\n\t}\n\tv191 := 0\n\tif 191 < len(data) {\n\t\tv191 = data[191]\n\t}\n\tv192 := 0\n\tif 192 < len(data) {\n\t\tv192 = data[192]\n\t}\n\tv193 := 0\n\tif 193 < len(data) {\n\t\tv193 = data[193]\n\t}\n\tv194 := 0\n\tif 194 < len(data) {\n\t\tv194 = data[194]\n\t}\n\tv195 := 0\n\tif 195 < len(data) {\n\t\tv195 = data[195]\n\t}\n\tv196 := 0\n\tif 196 < len(data) {\n\t\tv196 = data[196]\n\t}\n\tv197 := 0\n\tif 197 < len(data) {\n\t\tv197 = data[197]\n\t}\n\tv198 := 0\n\tif 198 < len(data) {\n\t\tv198 = data[198]\n\t}\n\tv199 := 0\n\tif 199 < len(data) {\n\t\tv199 = data[199]",
+    "token_estimate": 917
+  },
+  {
+    "block_ids": [
+      "5d269745b2e5dbdcbef0c09ba54b0bd6"
+    ],
+    "chunk_id": "438127626378632c03780d10603de32c",
+    "chunker_version": "code-go-ast-v1",
+    "doc_id": "83daba5fbb026e7a400d68a1c4bd36db",
+    "heading_path": [],
+    "policy_hash": "6cfe77abe2b0e5c3",
+    "source_spans": [
+      {
+        "kind": "code",
+        "lang": "go",
+        "line_end": 890,
+        "line_start": 848,
+        "symbol": "BigCompute [part 5/5]"
+      }
+    ],
+    "text": "\t}\n\tv200 := 0\n\tif 200 < len(data) {\n\t\tv200 = data[200]\n\t}\n\tv201 := 0\n\tif 201 < len(data) {\n\t\tv201 = data[201]\n\t}\n\tv202 := 0\n\tif 202 < len(data) {\n\t\tv202 = data[202]\n\t}\n\tv203 := 0\n\tif 203 < len(data) {\n\t\tv203 = data[203]\n\t}\n\tv204 := 0\n\tif 204 < len(data) {\n\t\tv204 = data[204]\n\t}\n\tv205 := 0\n\tif 205 < len(data) {\n\t\tv205 = data[205]\n\t}\n\tv206 := 0\n\tif 206 < len(data) {\n\t\tv206 = data[206]\n\t}\n\tv207 := 0\n\tif 207 < len(data) {\n\t\tv207 = data[207]\n\t}\n\tv208 := 0\n\tif 208 < len(data) {\n\t\tv208 = data[208]\n\t}\n\tv209 := 0\n\tif 209 < len(data) {\n\t\tv209 = data[209]\n\t}\n\treturn len(data)\n}",
+    "token_estimate": 191
+  }
+]

crates/kebab-cli/src/main.rs

@@ -275,6 +275,14 @@ enum Cmd {
         #[arg(long, group = "reset_scope")]
         config_only: bool,
 
+        /// Purge stored docs that are outside the current walker scope
+        /// (config narrowing / removed sub-directory). No filesystem paths
+        /// are removed — this is purely a store-level reconciliation.
+        /// Filesystem existence is NOT checked; anything the current walker
+        /// would not visit is considered an orphan and removed from the store.
+        #[arg(long, group = "reset_scope")]
+        orphans_only: bool,
+
         /// Skip the interactive confirm. Required in non-interactive
         /// contexts (CI, pipes).
         #[arg(long)]
@@ -595,14 +603,20 @@ fn run(cli: &Cli) -> anyhow::Result<()> {
                 println!("{}", serde_json::to_string(&wire::wire_ingest(&report))?);
             } else {
                 let skipped_breakdown = kebab_app::render_skipped_breakdown(&report.skipped_by_extension);
+                let purged_suffix = if report.purged_deleted_files > 0 {
+                    format!("  purged {}", report.purged_deleted_files)
+                } else {
+                    String::new()
+                };
                 println!(
-                    "scanned {}  new {}  updated {}  skipped {}{}  errors {}  ({} ms)",
+                    "scanned {}  new {}  updated {}  skipped {}{}  errors {}{}  ({} ms)",
                     report.scanned,
                     report.new,
                     report.updated,
                     report.skipped,
                     skipped_breakdown,
                     report.errors,
+                    purged_suffix,
                     report.duration_ms
                 );
             }
@@ -1088,6 +1102,7 @@ fn run(cli: &Cli) -> anyhow::Result<()> {
             data_only: _,
             vector_only,
             config_only,
+            orphans_only,
             yes,
         } => {
             use kebab_app::ResetScope;
@@ -1101,11 +1116,50 @@ fn run(cli: &Cli) -> anyhow::Result<()> {
                 ResetScope::VectorOnly
             } else if *config_only {
                 ResetScope::ConfigOnly
+            } else if *orphans_only {
+                ResetScope::OrphansOnly
             } else {
                 ResetScope::DataOnly
             };
 
             let cfg = kebab_config::Config::load(cli.config.as_deref())?;
+
+            if matches!(scope, ResetScope::OrphansOnly) {
+                // OrphansOnly: confirm UI shows orphan count + sample paths
+                // rather than on-disk directory sizes.
+                let orphan_paths = kebab_app::enumerate_orphans(&cfg)?;
+
+                if !*yes {
+                    use std::io::IsTerminal;
+                    if !std::io::stdin().is_terminal() {
+                        anyhow::bail!(
+                            "reset --orphans-only is destructive and stdin is non-interactive — pass --yes to proceed"
+                        );
+                    }
+                    if !confirm_orphans_only(&orphan_paths)? {
+                        if !cli.quiet {
+                            eprintln!("aborted.");
+                        }
+                        return Ok(());
+                    }
+                }
+
+                let report = kebab_app::reset::execute(scope, &cfg)?;
+                if cli.json {
+                    println!("{}", serde_json::to_string(&wire::wire_reset(&report))?);
+                } else {
+                    if report.orphans_purged > 0 {
+                        println!("orphans purged: {}", report.orphans_purged);
+                        for p in &report.purged_paths {
+                            println!("  - {}", p.0);
+                        }
+                    } else {
+                        println!("no orphaned docs found — store is already in sync with walker scope");
+                    }
+                }
+                return Ok(());
+            }
+
             let paths = kebab_app::reset::enumerate_paths(scope, &cfg);
             let bytes = kebab_app::reset::estimate_size_bytes(&paths);
 
@@ -1444,6 +1498,46 @@ fn confirm_destructive(
     Ok(matches!(s.as_str(), "y" | "yes"))
 }
 
+/// Confirm prompt for `--orphans-only`: shows the orphan count + a
+/// sample of up to 5 paths so the user knows what will be purged before
+/// committing. No filesystem paths are removed — only store records.
+fn confirm_orphans_only(
+    orphan_paths: &[kebab_core::WorkspacePath],
+) -> anyhow::Result<bool> {
+    use std::io::Write;
+    let n = orphan_paths.len();
+    let mut out = std::io::stderr().lock();
+
+    if n == 0 {
+        writeln!(out, "no orphaned docs found — nothing to purge.")?;
+        out.flush()?;
+        // Nothing to do; treat as confirmed so the caller can emit the
+        // "no orphans" report without prompting.
+        return Ok(true);
+    }
+
+    let sample: Vec<&str> = orphan_paths
+        .iter()
+        .take(5)
+        .map(|p| p.0.as_str())
+        .collect();
+    let sample_str = sample.join(", ");
+    let ellipsis = if n > 5 { ", …" } else { "" };
+
+    writeln!(
+        out,
+        "Purge {n} stored doc(s) outside the current walker scope? (no filesystem paths removed)"
+    )?;
+    writeln!(out, "  sample: {sample_str}{ellipsis}")?;
+    write!(out, "[y/N] ")?;
+    out.flush()?;
+
+    let mut line = String::new();
+    std::io::stdin().read_line(&mut line)?;
+    let s = line.trim().to_ascii_lowercase();
+    Ok(matches!(s.as_str(), "y" | "yes"))
+}
+
 /// p9-fb-35: human-friendly plain output for `kebab fetch`.
 fn render_fetch_plain(r: &kebab_core::FetchResult) {
     println!("# {} ({})", r.doc_path.0, format_kind(r.kind));

crates/kebab-cli/src/wire.rs

@@ -260,6 +260,7 @@ mod tests {
             skipped_generated: 0,
             skipped_size_exceeded: 0,
             skip_examples: SkipExamples::default(),
+            purged_deleted_files: 0,
             items: None,
         };
         let v = wire_ingest(&r);
@@ -364,6 +365,8 @@ mod tests {
             scope: kebab_app::ResetScope::DataOnly,
             removed_paths: vec![std::path::PathBuf::from("/tmp/x")],
             embedding_rows_truncated: 0,
+            orphans_purged: 0,
+            purged_paths: vec![],
         };
         let v = wire_reset(&r);
         assert_eq!(schema_of(&v), Some("reset_report.v1"));

crates/kebab-core/src/ingest.rs

@@ -47,6 +47,12 @@ pub struct IngestReport {
     /// p10-1A-1: sample file paths per skip category (≤ 5 each).
     #[serde(default)]
     pub skip_examples: SkipExamples,
+    /// Dogfood: docs whose on-disk file was deleted since the last ingest
+    /// and were therefore removed from the store. Additive field — older
+    /// wire consumers that pre-date this field read it as 0 via
+    /// `#[serde(default)]`.
+    #[serde(default)]
+    pub purged_deleted_files: u32,
     /// `None` ↔ wire `items: null` (`--summary-only`).
     pub items: Option<Vec<IngestItem>>,
 }
@@ -136,6 +142,7 @@ mod tests {
                 builtin_blacklist: vec!["node_modules/x.js".into()],
                 gitignore: vec![],
             },
+            purged_deleted_files: 0,
             items: None,
         };
         let v = serde_json::to_value(&r).unwrap();

crates/kebab-core/src/traits.rs

@@ -8,7 +8,7 @@ use serde_json::Value;
 use crate::asset::{RawAsset, WorkspacePath};
 use crate::chunk::Chunk;
 use crate::document::{Block, CanonicalDocument};
-use crate::ids::{ChunkId, DocumentId};
+use crate::ids::{AssetId, ChunkId, DocumentId};
 use crate::jobs::{JobFilter, JobId, JobKind, JobRow, JobStatus};
 use crate::media::MediaType;
 use crate::search::{DocFilter, DocSummary, SearchFilters, SearchHit, SearchQuery};
@@ -161,14 +161,51 @@ pub trait DocumentStore {
     fn get_document(&self, id: &DocumentId) -> anyhow::Result<Option<CanonicalDocument>>;
     fn get_chunk(&self, id: &ChunkId) -> anyhow::Result<Option<Chunk>>;
     fn list_documents(&self, filter: &DocFilter) -> anyhow::Result<Vec<DocSummary>>;
+    /// Look up an asset row by its `asset_id` (PRIMARY KEY = blake3
+    /// content hash). Twin-file safe: asset_id is PK so there is
+    /// exactly one row per unique content hash, regardless of how many
+    /// `documents` rows share it. Use this instead of
+    /// `get_asset_by_workspace_path` when you already have a
+    /// `CanonicalDocument` (which carries `source_asset_id`).
+    fn get_asset(&self, id: &AssetId) -> anyhow::Result<Option<RawAsset>>;
+
     /// p9-fb-23: look up an asset row by its workspace path. Used by
     /// the incremental-ingest skip path to compare the freshly
     /// computed blake3 checksum against what's already in SQLite. The
     /// schema enforces a unique workspace_path per asset.
+    ///
+    /// NOTE: for twin files (identical content at different paths),
+    /// `assets.workspace_path` is "last-registered path" — it
+    /// flip-flops on every ingest. Prefer `get_asset` (by asset_id)
+    /// when you have a `CanonicalDocument.source_asset_id`.
     fn get_asset_by_workspace_path(
         &self,
         path: &WorkspacePath,
     ) -> anyhow::Result<Option<RawAsset>>;
+
+    /// Look up a document row by its workspace path. Used by the
+    /// document-centric skip path in `try_skip_unchanged` to avoid the
+    /// twin-file flip-flop that the asset-side lookup suffers from
+    /// (multiple files with identical content share one `assets` row
+    /// whose `workspace_path` is overwritten on every UPSERT, so
+    /// `get_asset_by_workspace_path` returns the wrong twin's path).
+    ///
+    /// `documents.workspace_path` is UNIQUE (V001), so each twin has
+    /// its own stable document row regardless of the asset de-dup.
+    fn get_document_by_workspace_path(
+        &self,
+        path: &WorkspacePath,
+    ) -> anyhow::Result<Option<CanonicalDocument>>;
+
+    /// Return every `workspace_path` stored in the `documents` table.
+    ///
+    /// Used by the post-walker sweep in `kebab-app::ingest` to detect
+    /// documents whose source file has been deleted from the filesystem.
+    /// The set difference `(stored - scanned)` yields orphan candidates;
+    /// each candidate is then existence-checked on disk so that
+    /// out-of-scope files (config narrowing) are NOT purged — only truly
+    /// absent files trigger the purge.
+    fn all_workspace_paths(&self) -> anyhow::Result<Vec<WorkspacePath>>;
 }
 
 pub trait VectorStore {

crates/kebab-eval/tests/runner.rs

@@ -336,21 +336,29 @@ fn runner_lexical_is_deterministic_per_query_payload() {
         "- id: q1\n  query: ownership\n- id: q2\n  query: heading\n",
     );
 
-    let run_a = run_with_golden(&yaml, || {
+    let mut run_a = run_with_golden(&yaml, || {
         run_eval_with_config(&env.config, &lexical_opts()).unwrap()
     });
-    let run_b = run_with_golden(&yaml, || {
+    let mut run_b = run_with_golden(&yaml, || {
         run_eval_with_config(&env.config, &lexical_opts()).unwrap()
     });
 
     // Run-level fields (`run_id`, `created_at`) intentionally diverge;
     // the per-query payload (which is what the snapshot fixture pins)
-    // must be byte-identical.
+    // must be byte-identical EXCEPT for `elapsed_ms`. Timing-sensitive
+    // fields aren't determinism signals — they're µs-scale wall-clock
+    // jitter and would otherwise make this assertion a flaky one (a 0
+    // vs 1 ms divergence was observed under contended-CI load). Normalize
+    // before comparing; see test #7 for the same exclusion done via a
+    // projection.
+    for qr in run_a.per_query.iter_mut().chain(run_b.per_query.iter_mut()) {
+        qr.elapsed_ms = 0;
+    }
     let a_json = serde_json::to_string(&run_a.per_query).unwrap();
     let b_json = serde_json::to_string(&run_b.per_query).unwrap();
     assert_eq!(
         a_json, b_json,
-        "lexical-only per_query payload must be byte-identical across runs"
+        "lexical-only per_query payload must be byte-identical across runs (timing normalized)"
     );
 }
 

crates/kebab-parse-code/Cargo.toml

@@ -16,6 +16,12 @@ time             = { workspace = true }
 tracing          = { workspace = true }
 tree-sitter      = { workspace = true }
 tree-sitter-rust = { workspace = true }
+tree-sitter-python     = { workspace = true }
+tree-sitter-typescript = { workspace = true }
+tree-sitter-javascript = { workspace = true }
+tree-sitter-go         = { workspace = true }
+tree-sitter-java       = { workspace = true }
+tree-sitter-kotlin-ng  = { workspace = true }
 
 [dev-dependencies]
 tempfile = { workspace = true }

crates/kebab-parse-code/src/go.rs

@@ -0,0 +1,451 @@
+//! `kebab-parse-code::go` — tree-sitter Go AST extractor (P10-1C-Go Task D).
+//!
+//! Implements [`kebab_core::Extractor`] for [`MediaType::Code("go")`].
+//! Walks the tree-sitter parse tree and emits one [`Block::Code`] per
+//! top-level AST semantic unit (free fn, method, each type spec) carrying
+//! [`SourceSpan::Code`] with the unit's self-reference symbol path
+//! (design §3.4 Go row). Glue declarations (`import` / `const` / `var`)
+//! collapse into one grouped `<top-level>` (or `<module>`) unit.
+//!
+//! Unlike the Python/TS/JS extractors which path-derive their module
+//! prefix from the workspace file path, Go's package identity comes from
+//! the source itself (the leading `package` clause) — `extract_package`
+//! reads it from the AST. If the `package_clause` is missing (invalid Go
+//! in practice) the prefix falls back to `"<unknown>"`.
+//!
+//! Doc comments immediately preceding an item are folded into that
+//! item's line range via `unit_start` (1B pattern). Go has no separate
+//! attribute/decorator AST nodes.
+//!
+//! Per design §3.4 / §9.1 / §9 versioning.
+
+use anyhow::Result;
+use kebab_core::{
+    Block, CanonicalDocument, CodeBlock, CommonBlock, Extractor, Lang, MediaType, Metadata,
+    ParserVersion, Provenance, ProvenanceEvent, ProvenanceKind, SourceSpan, SourceType, TrustLevel,
+    id_for_block, id_for_doc,
+};
+use serde_json::Map;
+use time::OffsetDateTime;
+
+use crate::scaffold::{filename_from_workspace_path, join_symbol, strip_extension};
+
+pub const PARSER_VERSION: &str = "code-go-v1";
+
+/// Go AST extractor. Per-unit blocks via tree-sitter-go 0.25
+/// (`LANGUAGE: LanguageFn`) parsed by tree-sitter 0.26.
+pub struct GoAstExtractor;
+
+impl GoAstExtractor {
+    pub fn new() -> Self {
+        Self
+    }
+}
+
+impl Default for GoAstExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl Extractor for GoAstExtractor {
+    fn supports(&self, m: &MediaType) -> bool {
+        matches!(m, MediaType::Code(l) if l == "go")
+    }
+
+    fn parser_version(&self) -> ParserVersion {
+        ParserVersion(PARSER_VERSION.to_string())
+    }
+
+    fn extract(
+        &self,
+        ctx: &kebab_core::ExtractContext<'_>,
+        bytes: &[u8],
+    ) -> Result<CanonicalDocument> {
+        let asset = ctx.asset;
+        if !self.supports(&asset.media_type) {
+            anyhow::bail!(
+                "kebab-parse-code: unsupported media_type for GoAstExtractor: {:?}",
+                asset.media_type
+            );
+        }
+
+        let parser_version = self.parser_version();
+        let doc_id = id_for_doc(&asset.workspace_path, &asset.asset_id, &parser_version);
+
+        let source = String::from_utf8(bytes.to_vec())
+            .map_err(|e| anyhow::anyhow!("kebab-parse-code: Go source is not valid UTF-8: {e}"))?;
+
+        let blocks = build_blocks(&source, &doc_id)?;
+        let unit_count = blocks.len() as u32;
+
+        let now = OffsetDateTime::now_utc();
+        let mut events: Vec<ProvenanceEvent> = Vec::with_capacity(2);
+        events.push(ProvenanceEvent {
+            at: asset.discovered_at,
+            agent: "kb-source-fs".to_string(),
+            kind: ProvenanceKind::Discovered,
+            note: None,
+        });
+        events.push(ProvenanceEvent {
+            at: now,
+            agent: "kb-parse-code".to_string(),
+            kind: ProvenanceKind::Parsed,
+            note: Some(format!(
+                "parser_version={}; unit_count={}",
+                parser_version.0, unit_count
+            )),
+        });
+
+        let title = {
+            let fname = filename_from_workspace_path(&asset.workspace_path.0);
+            strip_extension(&fname)
+        };
+
+        // Resolve the file's absolute path for repo detection. If the
+        // source URI carries a relative path, anchor it at the workspace
+        // root so the `.git/` walk-up starts from the right place.
+        let abs_path = match &asset.source_uri {
+            kebab_core::SourceUri::File(p) => {
+                if p.is_absolute() {
+                    p.clone()
+                } else {
+                    ctx.workspace_root.join(p)
+                }
+            }
+            kebab_core::SourceUri::Kb(_) => ctx.workspace_root.to_path_buf(),
+        };
+        let (repo, git_branch, git_commit) = match crate::repo::detect_repo(&abs_path) {
+            Some(r) => (Some(r.name), r.branch, r.commit),
+            None => (None, None, None),
+        };
+
+        let metadata = Metadata {
+            aliases: Vec::new(),
+            tags: Vec::new(),
+            created_at: asset.discovered_at,
+            updated_at: asset.discovered_at,
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Map::new(),
+            repo,
+            git_branch,
+            git_commit,
+            code_lang: Some("go".to_string()),
+        };
+
+        tracing::debug!(
+            target: "kebab-parse-code",
+            "extracted Go doc_id={} workspace_path={} units={}",
+            doc_id.0,
+            asset.workspace_path.0,
+            unit_count
+        );
+
+        Ok(CanonicalDocument {
+            doc_id,
+            source_asset_id: asset.asset_id.clone(),
+            workspace_path: asset.workspace_path.clone(),
+            title,
+            lang: Lang("und".to_string()),
+            blocks,
+            metadata,
+            provenance: Provenance { events },
+            parser_version,
+            schema_version: 1,
+            doc_version: 1,
+            last_chunker_version: None,
+            last_embedding_version: None,
+        })
+    }
+}
+
+/// p10-1C-Go: extract `package` declaration text from a tree-sitter-go
+/// `source_file`. Returns `None` if no `package_clause` (invalid Go in
+/// practice but defense-in-depth). Per design §3.4 Go row.
+fn extract_package(root: tree_sitter::Node, src: &str) -> Option<String> {
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        if child.kind() == "package_clause" {
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                if sub.kind() == "package_identifier" {
+                    return Some(src[sub.start_byte()..sub.end_byte()].to_string());
+                }
+            }
+        }
+    }
+    None
+}
+
+fn build_blocks(
+    source: &str,
+    doc_id: &kebab_core::DocumentId,
+) -> anyhow::Result<Vec<kebab_core::Block>> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&tree_sitter_go::LANGUAGE.into())
+        .map_err(|e| anyhow::anyhow!("set tree-sitter-go language: {e}"))?;
+    let tree = parser
+        .parse(source.as_bytes(), None)
+        .ok_or_else(|| anyhow::anyhow!("tree-sitter failed to parse Go source"))?;
+    let lines: Vec<&str> = source.split('\n').collect();
+
+    let root = tree.root_node();
+    let mod_prefix = extract_package(root, source).unwrap_or_else(|| "<unknown>".to_string());
+
+    // units: (symbol, line_start, line_end, is_real_semantic_unit).
+    // Glue groups are pushed with a sentinel symbol + is_real=false so a
+    // post-pass can decide `<module>` vs `<top-level>` (1B post-pass
+    // mirror).
+    let mut units: Vec<(String, u32, u32, bool)> = Vec::new();
+    // (is_import 0/1, s, e). `is_import` flags `import_declaration` —
+    // used by the glue flush to pick `<module>` vs `<top-level>`
+    // provisional label.
+    let mut glue: Vec<(usize, u32, u32)> = Vec::new();
+
+    fn node_name_text<'a>(n: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+        n.child_by_field_name("name")
+            .map(|c| &src[c.start_byte()..c.end_byte()])
+    }
+    /// Walk preceding `comment` siblings to extend the unit's line range
+    /// upward, folding leading doc / line comments into the unit. Go has
+    /// no decorator/attribute nodes — doc comments are simply preceding
+    /// `comment` siblings (the 1B pattern).
+    fn unit_start(n: &tree_sitter::Node) -> u32 {
+        let mut start = n.start_position().row as u32 + 1;
+        let mut prev = n.prev_sibling();
+        while let Some(p) = prev {
+            if p.kind() == "comment" {
+                start = p.start_position().row as u32 + 1;
+                prev = p.prev_sibling();
+            } else {
+                break;
+            }
+        }
+        start
+    }
+
+    /// Extract the receiver type text for a `method_declaration`. The
+    /// returned slice INCLUDES the leading `*` for pointer receivers
+    /// (`(*Foo).Bar`) per design §3.4 Go row example. Returns `None` if
+    /// the receiver is malformed (defense in depth).
+    fn receiver_type_text<'a>(method_node: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+        let recv = method_node.child_by_field_name("receiver")?;
+        let mut cw = recv.walk();
+        for p in recv.named_children(&mut cw) {
+            if p.kind() == "parameter_declaration" {
+                if let Some(ty) = p.child_by_field_name("type") {
+                    return Some(&src[ty.start_byte()..ty.end_byte()]);
+                }
+            }
+        }
+        None
+    }
+
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        let s = unit_start(&child);
+        let e = child.end_position().row as u32 + 1;
+        match child.kind() {
+            "function_declaration" => {
+                if let Some(name) = node_name_text(&child, source) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(&mut glue, &mut units, &mod_prefix);
+                    let sym = join_symbol(&mod_prefix, &[], name);
+                    units.push((sym, s, e, true));
+                }
+            }
+            "method_declaration" => {
+                if let Some(name_node) = child.child_by_field_name("name") {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(&mut glue, &mut units, &mod_prefix);
+                    let owner = receiver_type_text(&child, source).unwrap_or("<unknown>");
+                    let method_name = &source[name_node.start_byte()..name_node.end_byte()];
+                    let sym = format!("{mod_prefix}.({owner}).{method_name}");
+                    units.push((sym, s, e, true));
+                }
+            }
+            "type_declaration" => {
+                // One unit per inner `type_spec`. Each type_spec gets
+                // the type_declaration's whole upward-folded `s` range
+                // start so doc comments are attached to the first spec;
+                // subsequent specs use their own start. Match 1B
+                // pattern: keep the outer `s` only when there's a single
+                // spec; otherwise use the spec's own start.
+                let mut tcur = child.walk();
+                let specs: Vec<tree_sitter::Node> = child
+                    .named_children(&mut tcur)
+                    .filter(|c| c.kind() == "type_spec")
+                    .collect();
+                let single = specs.len() == 1;
+                for spec in specs {
+                    let name_node = match spec.child_by_field_name("name") {
+                        Some(n) => n,
+                        None => continue,
+                    };
+                    let spec_s = if single {
+                        s
+                    } else {
+                        spec.start_position().row as u32 + 1
+                    };
+                    let spec_e = spec.end_position().row as u32 + 1;
+                    glue.retain(|(_, gs, _)| *gs < spec_s);
+                    flush_glue(&mut glue, &mut units, &mod_prefix);
+                    let name = &source[name_node.start_byte()..name_node.end_byte()];
+                    let sym = join_symbol(&mod_prefix, &[], name);
+                    units.push((sym, spec_s, spec_e, true));
+                }
+            }
+            "import_declaration" => {
+                glue.push((1, s, e));
+            }
+            "const_declaration" | "var_declaration" => {
+                glue.push((0, s, e));
+            }
+            _ => {}
+        }
+    }
+    flush_glue(&mut glue, &mut units, &mod_prefix);
+
+    // `<module>` is correct only when the file produced no real unit.
+    // Otherwise the import/const/var-only group becomes `<top-level>`
+    // (same post-pass as 1B). Match on the suffix so the demotion stays
+    // mod-prefix-agnostic.
+    let has_real_unit = units.iter().any(|(_, _, _, is_real)| *is_real);
+    if has_real_unit {
+        for (sym, _, _, is_real) in units.iter_mut() {
+            if !*is_real && sym.ends_with("<module>") {
+                let pre = &sym[..sym.len() - "<module>".len()];
+                *sym = format!("{pre}<top-level>");
+            }
+        }
+    }
+
+    let total_lines = lines.len() as u32;
+    let mut blocks = Vec::with_capacity(units.len());
+    for (ordinal, (symbol, ls, le, _is_real)) in units.into_iter().enumerate() {
+        let line_start = ls.max(1);
+        let line_end = le.min(total_lines.max(1));
+        let span = SourceSpan::Code {
+            line_start,
+            line_end,
+            symbol: Some(symbol),
+            lang: Some("go".to_string()),
+        };
+        let block_id = id_for_block(doc_id, "code", &[], ordinal as u32, &span);
+        let code = lines[(line_start as usize - 1)..=(line_end as usize - 1)].join("\n");
+        blocks.push(Block::Code(CodeBlock {
+            common: CommonBlock {
+                block_id,
+                heading_path: Vec::new(),
+                source_span: span,
+            },
+            lang: Some("go".to_string()),
+            code,
+        }));
+    }
+    Ok(blocks)
+}
+
+fn flush_glue(
+    glue: &mut Vec<(usize, u32, u32)>,
+    units: &mut Vec<(String, u32, u32, bool)>,
+    mod_prefix: &str,
+) {
+    if glue.is_empty() {
+        return;
+    }
+    let s = glue.iter().map(|(_, a, _)| *a).min().unwrap();
+    let e = glue.iter().map(|(_, _, b)| *b).max().unwrap();
+    // Provisional label: `<module>` only if the group is exclusively
+    // imports (1A's `only_mod_decls` analog). The post-pass demotes any
+    // `<module>` to `<top-level>` if the file produced any real unit.
+    let only_imports = glue.iter().all(|(is_import, _, _)| *is_import == 1);
+    let label = if only_imports { "<module>" } else { "<top-level>" };
+    units.push((join_symbol(mod_prefix, &[], label), s, e, false));
+    glue.clear();
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(concat!(
+            env!("CARGO_MANIFEST_DIR"),
+            "/tests/fixtures/sample.go"
+        ))
+        .unwrap();
+        // Reuse the cross-language test-support helper promoted in 1B.
+        let asset = crate::rust::tests_support::fixed_code_asset("crates/x/src/sample.go", "go");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        GoAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_go() {
+        let e = GoAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("go".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+
+    #[test]
+    fn go_units_match_design_3_4_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc
+            .blocks
+            .iter()
+            .filter_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, lang, .. } => {
+                        assert_eq!(lang.as_deref(), Some("go"));
+                        symbol.clone()
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .collect();
+        syms.sort();
+        assert!(syms.iter().any(|s| s == "chunk.Free"), "got {syms:?}");
+        assert!(syms.iter().any(|s| s == "chunk.init"), "got {syms:?}");
+        assert!(
+            syms.iter().any(|s| s == "chunk.MdHeadingV1Chunker"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "chunk.(*MdHeadingV1Chunker).ChunkDoc"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "chunk.(MdHeadingV1Chunker).Name2"),
+            "got {syms:?}"
+        );
+        assert!(syms.iter().any(|s| s == "chunk.Stringer"), "got {syms:?}");
+        // import + const grouped into one glue unit (no isolated `<module>`).
+        assert!(
+            syms.iter().any(|s| s == "chunk.<top-level>"),
+            "got {syms:?}"
+        );
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 {
+            assert_eq!(extract_fixture().blocks, a.blocks);
+        }
+    }
+}

crates/kebab-parse-code/src/java.rs

@@ -0,0 +1,543 @@
+//! `kebab-parse-code::java` — tree-sitter Java AST extractor (P10-1C-JK Task D).
+//!
+//! Implements [`kebab_core::Extractor`] for [`MediaType::Code("java")`].
+//! Walks the tree-sitter parse tree and emits one [`Block::Code`] per
+//! top-level AST semantic unit (class / interface / enum / record /
+//! annotation-type at any nesting level, plus methods + constructors
+//! inside class / interface / record bodies), each carrying
+//! [`SourceSpan::Code`] with the unit's dotted self-reference symbol
+//! path (design §3.4 Java row). Glue declarations (`import`) collapse
+//! into one grouped `<top-level>` (or `<module>`) unit.
+//!
+//! Like the Go extractor, Java's package identity comes from the
+//! source itself (the `package_declaration` clause), not from the
+//! workspace file path — `extract_package` reads it from the AST. If
+//! the clause is missing the prefix falls back to `"<unknown>"`.
+//!
+//! Class/interface/record bodies are recursed (1B Python pattern):
+//! the type name is pushed onto `mod_path` so methods and nested
+//! types become `<pkg>.<Outer>.<Inner>.<method>`. Constructors use
+//! the Java convention `<pkg>.<...>.<Class>.<ClassName>` (name
+//! duplicated, per design §3.4). Enum bodies are not recursed for
+//! the 1차 cut — enum constants are not emitted as units.
+//!
+//! Javadoc (`/** ... */` → `block_comment`) and line comments
+//! immediately preceding an item are folded into that item's line
+//! range via `unit_start` (1B pattern). Annotations are children of
+//! the declaration node itself (inside `modifiers`), so they are
+//! already part of the declaration's span — no separate unwrap arm.
+//!
+//! Per design §3.4 / §9.1 / §9 versioning.
+
+use anyhow::Result;
+use kebab_core::{
+    Block, CanonicalDocument, CodeBlock, CommonBlock, Extractor, Lang, MediaType, Metadata,
+    ParserVersion, Provenance, ProvenanceEvent, ProvenanceKind, SourceSpan, SourceType, TrustLevel,
+    id_for_block, id_for_doc,
+};
+use serde_json::Map;
+use time::OffsetDateTime;
+
+use crate::scaffold::{filename_from_workspace_path, join_symbol, strip_extension};
+
+pub const PARSER_VERSION: &str = "code-java-v1";
+
+/// Java AST extractor. Per-unit blocks via tree-sitter-java 0.23
+/// (`LANGUAGE: LanguageFn`) parsed by tree-sitter 0.26.
+pub struct JavaAstExtractor;
+
+impl JavaAstExtractor {
+    pub fn new() -> Self {
+        Self
+    }
+}
+
+impl Default for JavaAstExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl Extractor for JavaAstExtractor {
+    fn supports(&self, m: &MediaType) -> bool {
+        matches!(m, MediaType::Code(l) if l == "java")
+    }
+
+    fn parser_version(&self) -> ParserVersion {
+        ParserVersion(PARSER_VERSION.to_string())
+    }
+
+    fn extract(
+        &self,
+        ctx: &kebab_core::ExtractContext<'_>,
+        bytes: &[u8],
+    ) -> Result<CanonicalDocument> {
+        let asset = ctx.asset;
+        if !self.supports(&asset.media_type) {
+            anyhow::bail!(
+                "kebab-parse-code: unsupported media_type for JavaAstExtractor: {:?}",
+                asset.media_type
+            );
+        }
+
+        let parser_version = self.parser_version();
+        let doc_id = id_for_doc(&asset.workspace_path, &asset.asset_id, &parser_version);
+
+        let source = String::from_utf8(bytes.to_vec())
+            .map_err(|e| anyhow::anyhow!("kebab-parse-code: Java source is not valid UTF-8: {e}"))?;
+
+        let blocks = build_blocks(&source, &doc_id)?;
+        let unit_count = blocks.len() as u32;
+
+        let now = OffsetDateTime::now_utc();
+        let mut events: Vec<ProvenanceEvent> = Vec::with_capacity(2);
+        events.push(ProvenanceEvent {
+            at: asset.discovered_at,
+            agent: "kb-source-fs".to_string(),
+            kind: ProvenanceKind::Discovered,
+            note: None,
+        });
+        events.push(ProvenanceEvent {
+            at: now,
+            agent: "kb-parse-code".to_string(),
+            kind: ProvenanceKind::Parsed,
+            note: Some(format!(
+                "parser_version={}; unit_count={}",
+                parser_version.0, unit_count
+            )),
+        });
+
+        let title = {
+            let fname = filename_from_workspace_path(&asset.workspace_path.0);
+            strip_extension(&fname)
+        };
+
+        // Resolve the file's absolute path for repo detection. If the
+        // source URI carries a relative path, anchor it at the workspace
+        // root so the `.git/` walk-up starts from the right place.
+        let abs_path = match &asset.source_uri {
+            kebab_core::SourceUri::File(p) => {
+                if p.is_absolute() {
+                    p.clone()
+                } else {
+                    ctx.workspace_root.join(p)
+                }
+            }
+            kebab_core::SourceUri::Kb(_) => ctx.workspace_root.to_path_buf(),
+        };
+        let (repo, git_branch, git_commit) = match crate::repo::detect_repo(&abs_path) {
+            Some(r) => (Some(r.name), r.branch, r.commit),
+            None => (None, None, None),
+        };
+
+        let metadata = Metadata {
+            aliases: Vec::new(),
+            tags: Vec::new(),
+            created_at: asset.discovered_at,
+            updated_at: asset.discovered_at,
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Map::new(),
+            repo,
+            git_branch,
+            git_commit,
+            code_lang: Some("java".to_string()),
+        };
+
+        tracing::debug!(
+            target: "kebab-parse-code",
+            "extracted Java doc_id={} workspace_path={} units={}",
+            doc_id.0,
+            asset.workspace_path.0,
+            unit_count
+        );
+
+        Ok(CanonicalDocument {
+            doc_id,
+            source_asset_id: asset.asset_id.clone(),
+            workspace_path: asset.workspace_path.clone(),
+            title,
+            lang: Lang("und".to_string()),
+            blocks,
+            metadata,
+            provenance: Provenance { events },
+            parser_version,
+            schema_version: 1,
+            doc_version: 1,
+            last_chunker_version: None,
+            last_embedding_version: None,
+        })
+    }
+}
+
+/// p10-1C-JK: extract `package` declaration text from a tree-sitter-java
+/// `program`. Returns `None` if no `package_declaration` (default-package
+/// Java file). The package_declaration's named children are either a
+/// single `identifier` (single-segment package, rare) or a
+/// `scoped_identifier` (dotted, common). Per design §3.4 Java row.
+fn extract_package(root: tree_sitter::Node, src: &str) -> Option<String> {
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        if child.kind() == "package_declaration" {
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                if sub.kind() == "scoped_identifier" || sub.kind() == "identifier" {
+                    return Some(src[sub.start_byte()..sub.end_byte()].to_string());
+                }
+            }
+        }
+    }
+    None
+}
+
+/// Walk preceding `line_comment` / `block_comment` siblings to extend
+/// the unit's line range upward, folding leading Javadoc / line
+/// comments into the unit. Annotations live INSIDE `modifiers` on the
+/// declaration node itself, so their lines are already inside
+/// `n.start_position()` — no separate unwrap arm is needed for them.
+fn unit_start(n: &tree_sitter::Node) -> u32 {
+    let mut start = n.start_position().row as u32 + 1;
+    let mut prev = n.prev_sibling();
+    while let Some(p) = prev {
+        let k = p.kind();
+        if k == "line_comment" || k == "block_comment" {
+            start = p.start_position().row as u32 + 1;
+            prev = p.prev_sibling();
+        } else {
+            break;
+        }
+    }
+    start
+}
+
+fn node_name_text<'a>(n: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+    n.child_by_field_name("name")
+        .map(|c| &src[c.start_byte()..c.end_byte()])
+}
+
+fn build_blocks(
+    source: &str,
+    doc_id: &kebab_core::DocumentId,
+) -> anyhow::Result<Vec<kebab_core::Block>> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&tree_sitter_java::LANGUAGE.into())
+        .map_err(|e| anyhow::anyhow!("set tree-sitter-java language: {e}"))?;
+    let tree = parser
+        .parse(source.as_bytes(), None)
+        .ok_or_else(|| anyhow::anyhow!("tree-sitter failed to parse Java source"))?;
+    let lines: Vec<&str> = source.split('\n').collect();
+
+    let root = tree.root_node();
+    let mod_prefix = extract_package(root, source).unwrap_or_else(|| "<unknown>".to_string());
+
+    // units: (symbol, line_start, line_end, is_real_semantic_unit).
+    // Glue groups are pushed with a sentinel symbol + is_real=false so a
+    // post-pass can decide `<module>` vs `<top-level>` (1B/1C-Go pattern).
+    let mut units: Vec<(String, u32, u32, bool)> = Vec::new();
+    // (is_import 0/1, s, e). `is_import` flags `import_declaration` —
+    // used by the glue flush to pick `<module>` vs `<top-level>`
+    // provisional label.
+    let mut glue: Vec<(usize, u32, u32)> = Vec::new();
+
+    walk_top(root, source, &mod_prefix, &mut units, &mut glue);
+
+    // `<module>` is correct only when the file produced no real unit.
+    // Otherwise the import-only group becomes `<top-level>` (same
+    // post-pass as 1B / 1C-Go).
+    let has_real_unit = units.iter().any(|(_, _, _, is_real)| *is_real);
+    if has_real_unit {
+        for (sym, _, _, is_real) in units.iter_mut() {
+            if !*is_real && sym.ends_with("<module>") {
+                let pre = &sym[..sym.len() - "<module>".len()];
+                *sym = format!("{pre}<top-level>");
+            }
+        }
+    }
+
+    let total_lines = lines.len() as u32;
+    let mut blocks = Vec::with_capacity(units.len());
+    for (ordinal, (symbol, ls, le, _is_real)) in units.into_iter().enumerate() {
+        let line_start = ls.max(1);
+        let line_end = le.min(total_lines.max(1));
+        let span = SourceSpan::Code {
+            line_start,
+            line_end,
+            symbol: Some(symbol),
+            lang: Some("java".to_string()),
+        };
+        let block_id = id_for_block(doc_id, "code", &[], ordinal as u32, &span);
+        let code = lines[(line_start as usize - 1)..=(line_end as usize - 1)].join("\n");
+        blocks.push(Block::Code(CodeBlock {
+            common: CommonBlock {
+                block_id,
+                heading_path: Vec::new(),
+                source_span: span,
+            },
+            lang: Some("java".to_string()),
+            code,
+        }));
+    }
+    Ok(blocks)
+}
+
+/// Walk the file's top-level children — `program` named children:
+/// `package_declaration` (handled by `extract_package`), `import_declaration`
+/// (glue), and the five type declarations (`class` / `interface` /
+/// `enum` / `record` / `annotation_type`). Type-declaration bodies
+/// are recursed via [`walk_body`] with the type name pushed onto
+/// `mod_path` (1B Python pattern). Enum bodies are NOT recursed
+/// (1차 cut — see module-level doc).
+fn walk_top(
+    node: tree_sitter::Node,
+    src: &str,
+    mod_prefix: &str,
+    units: &mut Vec<(String, u32, u32, bool)>,
+    glue: &mut Vec<(usize, u32, u32)>,
+) {
+    let mod_path: &[String] = &[];
+    let mut cur = node.walk();
+    for child in node.named_children(&mut cur) {
+        let s = unit_start(&child);
+        let e = child.end_position().row as u32 + 1;
+        match child.kind() {
+            "class_declaration"
+            | "interface_declaration"
+            | "record_declaration" => {
+                if let Some(name) = node_name_text(&child, src) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(glue, units, mod_prefix, mod_path);
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                    if let Some(body) = child.child_by_field_name("body") {
+                        let np: Vec<String> = vec![name.to_string()];
+                        walk_body(body, src, mod_prefix, &np, units);
+                    }
+                }
+            }
+            "enum_declaration" => {
+                if let Some(name) = node_name_text(&child, src) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(glue, units, mod_prefix, mod_path);
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                    // Enum body NOT recursed for 1차 — enum constants are
+                    // not emitted as units, and method declarations inside
+                    // enum bodies (rare) live under `enum_body_declarations`
+                    // not `class_body`. Skip per design §3.4 1차 scope.
+                }
+            }
+            "annotation_type_declaration" => {
+                if let Some(name) = node_name_text(&child, src) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(glue, units, mod_prefix, mod_path);
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                }
+            }
+            "import_declaration" => {
+                glue.push((1, s, e));
+            }
+            // package_declaration is handled by `extract_package`; no
+            // glue entry — it's structural metadata, not a unit.
+            _ => {}
+        }
+    }
+    flush_glue(glue, units, mod_prefix, mod_path);
+}
+
+/// Walk a `class_body` / `interface_body` (or record's `class_body`).
+/// Emits one unit per method / constructor, and recurses into nested
+/// type declarations. Field declarations are NOT emitted (would
+/// explode unit count). `compact_constructor_declaration` (records)
+/// is handled the same as `constructor_declaration`.
+///
+/// No `glue` parameter: Java does not have imports inside type
+/// bodies — they only appear at file top level, handled by
+/// [`walk_top`].
+fn walk_body(
+    body: tree_sitter::Node,
+    src: &str,
+    mod_prefix: &str,
+    mod_path: &[String],
+    units: &mut Vec<(String, u32, u32, bool)>,
+) {
+    let mut cur = body.walk();
+    for child in body.named_children(&mut cur) {
+        let s = unit_start(&child);
+        let e = child.end_position().row as u32 + 1;
+        match child.kind() {
+            "method_declaration"
+            | "constructor_declaration"
+            | "compact_constructor_declaration" => {
+                // Constructor: name field equals the class name. Per
+                // design §3.4 Java convention, symbol is
+                // `<pkg>.<mod_path>.<ClassName>` with the constructor
+                // name (== class name) as the trailing segment. This
+                // means the symbol duplicates the class name (e.g.
+                // `com.x.Foo.Foo`), which is the documented convention.
+                if let Some(name) = node_name_text(&child, src) {
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                }
+            }
+            "class_declaration"
+            | "interface_declaration"
+            | "record_declaration"
+            | "enum_declaration"
+            | "annotation_type_declaration" => {
+                // Nested type — emit unit, then recurse into its body
+                // (skipped for enum + annotation_type per 1차 scope).
+                let name = match node_name_text(&child, src) {
+                    Some(n) => n,
+                    None => continue,
+                };
+                let sym = join_symbol(mod_prefix, mod_path, name);
+                units.push((sym, s, e, true));
+                if child.kind() != "enum_declaration"
+                    && child.kind() != "annotation_type_declaration"
+                {
+                    if let Some(inner_body) = child.child_by_field_name("body") {
+                        let mut np = mod_path.to_vec();
+                        np.push(name.to_string());
+                        walk_body(inner_body, src, mod_prefix, &np, units);
+                    }
+                }
+            }
+            // field_declaration, static_initializer, block: NOT emitted.
+            _ => {}
+        }
+    }
+}
+
+fn flush_glue(
+    glue: &mut Vec<(usize, u32, u32)>,
+    units: &mut Vec<(String, u32, u32, bool)>,
+    mod_prefix: &str,
+    mod_path: &[String],
+) {
+    if glue.is_empty() {
+        return;
+    }
+    let s = glue.iter().map(|(_, a, _)| *a).min().unwrap();
+    let e = glue.iter().map(|(_, _, b)| *b).max().unwrap();
+    // Provisional label: `<module>` only if the group is exclusively
+    // imports (1A's `only_mod_decls` analog). The post-pass demotes any
+    // `<module>` to `<top-level>` if the file produced any real unit.
+    let only_imports = glue.iter().all(|(is_import, _, _)| *is_import == 1);
+    let label = if only_imports { "<module>" } else { "<top-level>" };
+    units.push((join_symbol(mod_prefix, mod_path, label), s, e, false));
+    glue.clear();
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(concat!(
+            env!("CARGO_MANIFEST_DIR"),
+            "/tests/fixtures/sample.java"
+        ))
+        .unwrap();
+        let asset =
+            crate::rust::tests_support::fixed_code_asset("crates/x/src/sample.java", "java");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        JavaAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_java() {
+        let e = JavaAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("java".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+
+    #[test]
+    fn java_units_match_design_3_4_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc
+            .blocks
+            .iter()
+            .filter_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, lang, .. } => {
+                        assert_eq!(lang.as_deref(), Some("java"));
+                        symbol.clone()
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .collect();
+        syms.sort();
+        // package extracted from source = com.kebab.chunk
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker"),
+            "got {syms:?}"
+        );
+        // constructor — Java convention is class-name-as-method-name
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.MdHeadingV1Chunker"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.chunkDoc"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.getName"),
+            "got {syms:?}"
+        );
+        // static nested class
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Builder"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Builder.withName"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Builder.build"),
+            "got {syms:?}"
+        );
+        // package-private interface + enum
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.Stringer"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.Mode"),
+            "got {syms:?}"
+        );
+        // import grouped as <top-level>
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.<top-level>"),
+            "got {syms:?}"
+        );
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 {
+            assert_eq!(extract_fixture().blocks, a.blocks);
+        }
+    }
+}

crates/kebab-parse-code/src/javascript.rs

@@ -0,0 +1,574 @@
+//! `kebab-parse-code::javascript` — tree-sitter JavaScript / JSX AST
+//! extractor (P10-1B Task K).
+//!
+//! Implements [`kebab_core::Extractor`] for [`MediaType::Code("javascript")`].
+//! Walks the tree-sitter parse tree (single grammar
+//! [`tree_sitter_javascript::LANGUAGE`] — the JS grammar handles `.jsx`
+//! as well, no second grammar needed) and emits one [`Block::Code`] per
+//! top-level AST semantic unit (free fn, class, each method,
+//! recursively per nested class), each carrying [`SourceSpan::Code`]
+//! with the unit's dotted symbol path prefixed by
+//! [`module_path_for_tsjs`].
+//!
+//! Glue declarations (`import_statement`, bare `export_statement`
+//! re-exports, `lexical_declaration` / `variable_declaration` at the
+//! module level, etc.) collapse into one grouped `<top-level>` (or
+//! `<module>`) unit.
+//!
+//! `export_statement` is unwrapped: an `export function|class` is
+//! treated as the inner declaration arm but the unit's line range
+//! comes from the OUTER `export_statement` so the `export ` prefix is
+//! folded in. `export default function () {}` / `export default class
+//! {}` (no `name` field) emits `default` as the symbol name.
+//!
+//! Differs from `typescript.rs` only by: single-grammar (no
+//! TS/TSX selection) and no `interface_declaration` /
+//! `type_alias_declaration` / `enum_declaration` arms (TS-only). All
+//! other walker behavior (export unwrap with `value`-field quirk for
+//! default-exported anonymous function/class, class-body method walk,
+//! glue flush, post-pass `<module>` → `<top-level>` rewrite) is
+//! identical.
+//!
+//! Scope follows 1A-2 / 1B Task K: AST unit extraction + dotted symbol
+//! paths + line ranges. Per design §3.4 / §9.1 / §9 versioning.
+
+use anyhow::Result;
+use kebab_core::{
+    Block, CanonicalDocument, CodeBlock, CommonBlock, Extractor, Lang, MediaType, Metadata,
+    ParserVersion, Provenance, ProvenanceEvent, ProvenanceKind, SourceSpan, SourceType, TrustLevel,
+    id_for_block, id_for_doc,
+};
+use serde_json::Map;
+use time::OffsetDateTime;
+
+use crate::scaffold::{filename_from_workspace_path, join_symbol, strip_extension};
+
+pub const PARSER_VERSION: &str = "code-js-v1";
+
+/// JavaScript / JSX AST extractor. Per-unit blocks via
+/// tree-sitter-javascript 0.25 (single `LANGUAGE` `LanguageFn` — the
+/// JS grammar covers `.jsx` natively, no second grammar) parsed by
+/// tree-sitter 0.26.
+pub struct JavascriptAstExtractor;
+
+impl JavascriptAstExtractor {
+    pub fn new() -> Self {
+        Self
+    }
+}
+
+impl Default for JavascriptAstExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl Extractor for JavascriptAstExtractor {
+    fn supports(&self, m: &MediaType) -> bool {
+        matches!(m, MediaType::Code(l) if l == "javascript")
+    }
+
+    fn parser_version(&self) -> ParserVersion {
+        ParserVersion(PARSER_VERSION.to_string())
+    }
+
+    fn extract(
+        &self,
+        ctx: &kebab_core::ExtractContext<'_>,
+        bytes: &[u8],
+    ) -> Result<CanonicalDocument> {
+        let asset = ctx.asset;
+        if !self.supports(&asset.media_type) {
+            anyhow::bail!(
+                "kebab-parse-code: unsupported media_type for JavascriptAstExtractor: {:?}",
+                asset.media_type
+            );
+        }
+
+        let parser_version = self.parser_version();
+        let doc_id = id_for_doc(&asset.workspace_path, &asset.asset_id, &parser_version);
+
+        let source = String::from_utf8(bytes.to_vec()).map_err(|e| {
+            anyhow::anyhow!("kebab-parse-code: JavaScript source is not valid UTF-8: {e}")
+        })?;
+
+        let mod_prefix = crate::lang::module_path_for_tsjs(&asset.workspace_path.0);
+        let language: tree_sitter::Language = tree_sitter_javascript::LANGUAGE.into();
+        let blocks = build_blocks(&source, &doc_id, &mod_prefix, language)?;
+        let unit_count = blocks.len() as u32;
+
+        let now = OffsetDateTime::now_utc();
+        let mut events: Vec<ProvenanceEvent> = Vec::with_capacity(2);
+        events.push(ProvenanceEvent {
+            at: asset.discovered_at,
+            agent: "kb-source-fs".to_string(),
+            kind: ProvenanceKind::Discovered,
+            note: None,
+        });
+        events.push(ProvenanceEvent {
+            at: now,
+            agent: "kb-parse-code".to_string(),
+            kind: ProvenanceKind::Parsed,
+            note: Some(format!(
+                "parser_version={}; unit_count={}",
+                parser_version.0, unit_count
+            )),
+        });
+
+        let title = {
+            let fname = filename_from_workspace_path(&asset.workspace_path.0);
+            strip_extension(&fname)
+        };
+
+        // Resolve the file's absolute path for repo detection. If the
+        // source URI carries a relative path, anchor it at the workspace
+        // root so the `.git/` walk-up starts from the right place.
+        let abs_path = match &asset.source_uri {
+            kebab_core::SourceUri::File(p) => {
+                if p.is_absolute() {
+                    p.clone()
+                } else {
+                    ctx.workspace_root.join(p)
+                }
+            }
+            kebab_core::SourceUri::Kb(_) => ctx.workspace_root.to_path_buf(),
+        };
+        let (repo, git_branch, git_commit) = match crate::repo::detect_repo(&abs_path) {
+            Some(r) => (Some(r.name), r.branch, r.commit),
+            None => (None, None, None),
+        };
+
+        let metadata = Metadata {
+            aliases: Vec::new(),
+            tags: Vec::new(),
+            created_at: asset.discovered_at,
+            updated_at: asset.discovered_at,
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Map::new(),
+            repo,
+            git_branch,
+            git_commit,
+            code_lang: Some("javascript".to_string()),
+        };
+
+        tracing::debug!(
+            target: "kebab-parse-code",
+            "extracted JavaScript doc_id={} workspace_path={} units={}",
+            doc_id.0,
+            asset.workspace_path.0,
+            unit_count
+        );
+
+        Ok(CanonicalDocument {
+            doc_id,
+            source_asset_id: asset.asset_id.clone(),
+            workspace_path: asset.workspace_path.clone(),
+            title,
+            lang: Lang("und".to_string()),
+            blocks,
+            metadata,
+            provenance: Provenance { events },
+            parser_version,
+            schema_version: 1,
+            doc_version: 1,
+            last_chunker_version: None,
+            last_embedding_version: None,
+        })
+    }
+}
+
+fn build_blocks(
+    source: &str,
+    doc_id: &kebab_core::DocumentId,
+    mod_prefix: &str,
+    language: tree_sitter::Language,
+) -> anyhow::Result<Vec<kebab_core::Block>> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&language)
+        .map_err(|e| anyhow::anyhow!("set tree-sitter-javascript language: {e}"))?;
+    let tree = parser
+        .parse(source.as_bytes(), None)
+        .ok_or_else(|| anyhow::anyhow!("tree-sitter failed to parse JavaScript source"))?;
+    let lines: Vec<&str> = source.split('\n').collect();
+
+    // units: (symbol, line_start, line_end, is_real_semantic_unit).
+    // Glue groups are pushed with a sentinel symbol + is_real=false so a
+    // post-pass can decide `<module>` vs `<top-level>` (same algorithm
+    // as 1A Gap 1 / 1B Python / 1B TS).
+    let mut units: Vec<(String, u32, u32, bool)> = Vec::new();
+    // (is_module_only_kind 0/1, s, e). `is_module_only_kind` flags
+    // `import_statement` and bare re-export `export_statement`s — used by
+    // the glue flush to pick `<module>` vs `<top-level>` provisional
+    // label (1A's `is_mod_decl` analog).
+    let mut glue: Vec<(usize, u32, u32)> = Vec::new();
+
+    /// Walk preceding `comment` siblings to extend the unit's line range
+    /// upward, folding leading doc / line comments into the unit.
+    fn unit_start(n: &tree_sitter::Node) -> u32 {
+        let mut start = n.start_position().row as u32 + 1;
+        let mut prev = n.prev_sibling();
+        while let Some(p) = prev {
+            if p.kind() == "comment" {
+                start = p.start_position().row as u32 + 1;
+                prev = p.prev_sibling();
+            } else {
+                break;
+            }
+        }
+        start
+    }
+    fn name_text<'a>(n: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+        n.child_by_field_name("name")
+            .map(|c| &src[c.start_byte()..c.end_byte()])
+    }
+    /// Walk a class body, emitting one unit per `method_definition`.
+    /// Class names already pushed onto `mod_path` by the caller, so
+    /// method symbols come out as `<mod_prefix>.<Class>.<method>`.
+    fn walk_class_body(
+        body: tree_sitter::Node,
+        src: &str,
+        mod_prefix: &str,
+        mod_path: &[String],
+        units: &mut Vec<(String, u32, u32, bool)>,
+    ) {
+        let mut cur = body.walk();
+        for child in body.named_children(&mut cur) {
+            if child.kind() == "method_definition" {
+                if let Some(name) = name_text(&child, src) {
+                    let s = unit_start(&child);
+                    let e = child.end_position().row as u32 + 1;
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                }
+            }
+        }
+    }
+    fn walk(
+        node: tree_sitter::Node,
+        src: &str,
+        mod_prefix: &str,
+        mod_path: &[String],
+        units: &mut Vec<(String, u32, u32, bool)>,
+        glue: &mut Vec<(usize, u32, u32)>,
+    ) {
+        let mut cur = node.walk();
+        for child in node.named_children(&mut cur) {
+            let s = unit_start(&child);
+            let e = child.end_position().row as u32 + 1;
+            match child.kind() {
+                "function_declaration" => {
+                    if let Some(name) = name_text(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                    }
+                }
+                "class_declaration" => {
+                    if let Some(name) = name_text(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                        if let Some(body) = child.child_by_field_name("body") {
+                            let mut np = mod_path.to_vec();
+                            np.push(name.to_string());
+                            walk_class_body(body, src, mod_prefix, &np, units);
+                        }
+                    }
+                }
+                "export_statement" => {
+                    // Try field "declaration" first (export class /
+                    // function). If absent, fall back to "value" —
+                    // `export default function () {}` / `export default
+                    // class {}` expose the anonymous function_expression
+                    // / class under the `value` field (same grammar
+                    // quirk as TS 0.23).
+                    let outer_s = s; // includes `export ` prefix line
+                    let outer_e = e;
+                    if let Some(inner) = child.child_by_field_name("declaration") {
+                        let inner_kind = inner.kind();
+                        match inner_kind {
+                            "function_declaration" | "class_declaration" => {
+                                let name_opt = name_text(&inner, src).map(|s| s.to_string());
+                                if let Some(name) = name_opt {
+                                    glue.retain(|(_, gs, _)| *gs < outer_s);
+                                    flush_glue(glue, units, mod_prefix, mod_path);
+                                    let sym = join_symbol(mod_prefix, mod_path, &name);
+                                    units.push((sym, outer_s, outer_e, true));
+                                    if inner_kind == "class_declaration" {
+                                        if let Some(body) = inner.child_by_field_name("body") {
+                                            let mut np = mod_path.to_vec();
+                                            np.push(name);
+                                            walk_class_body(body, src, mod_prefix, &np, units);
+                                        }
+                                    }
+                                } else {
+                                    // Defensive: `export default` with a
+                                    // function_declaration that somehow
+                                    // lacks `name`. Emit `default`.
+                                    glue.retain(|(_, gs, _)| *gs < outer_s);
+                                    flush_glue(glue, units, mod_prefix, mod_path);
+                                    let sym = join_symbol(mod_prefix, mod_path, "default");
+                                    units.push((sym, outer_s, outer_e, true));
+                                }
+                            }
+                            // `lexical_declaration` etc. wrapped in
+                            // export: treat as glue (assigned arrow
+                            // fns / consts don't get their own unit).
+                            _ => {
+                                glue.push((0, s, e));
+                            }
+                        }
+                    } else if let Some(value) = child.child_by_field_name("value") {
+                        // `export default <expr>`. We emit a unit only
+                        // for the function / class shapes (named or
+                        // anonymous); other value shapes are glue.
+                        match value.kind() {
+                            "function_expression"
+                            | "function_declaration"
+                            | "class"
+                            | "class_declaration" => {
+                                let name_opt = name_text(&value, src).map(|s| s.to_string());
+                                let leaf =
+                                    name_opt.as_deref().unwrap_or("default").to_string();
+                                glue.retain(|(_, gs, _)| *gs < outer_s);
+                                flush_glue(glue, units, mod_prefix, mod_path);
+                                let sym = join_symbol(mod_prefix, mod_path, &leaf);
+                                units.push((sym, outer_s, outer_e, true));
+                                // Recurse into class body if we have one.
+                                if matches!(value.kind(), "class" | "class_declaration") {
+                                    if let Some(body) = value.child_by_field_name("body") {
+                                        let mut np = mod_path.to_vec();
+                                        np.push(leaf);
+                                        walk_class_body(body, src, mod_prefix, &np, units);
+                                    }
+                                }
+                            }
+                            _ => {
+                                glue.push((0, s, e));
+                            }
+                        }
+                    } else {
+                        // Bare `export { x };` / `export * from "..."` —
+                        // a re-export, glue with module-only flag set
+                        // (we have no `declaration` / `value` field for
+                        // it).
+                        glue.push((1, s, e));
+                    }
+                }
+                "import_statement" => {
+                    glue.push((1, s, e));
+                }
+                "lexical_declaration" | "variable_declaration" => {
+                    glue.push((0, s, e));
+                }
+                _ => {}
+            }
+        }
+        flush_glue(glue, units, mod_prefix, mod_path);
+    }
+    fn flush_glue(
+        glue: &mut Vec<(usize, u32, u32)>,
+        units: &mut Vec<(String, u32, u32, bool)>,
+        mod_prefix: &str,
+        mod_path: &[String],
+    ) {
+        if glue.is_empty() {
+            return;
+        }
+        let s = glue.iter().map(|(_, a, _)| *a).min().unwrap();
+        let e = glue.iter().map(|(_, _, b)| *b).max().unwrap();
+        let only_module = glue.iter().all(|(is_mod, _, _)| *is_mod == 1);
+        let label = if only_module { "<module>" } else { "<top-level>" };
+        units.push((join_symbol(mod_prefix, mod_path, label), s, e, false));
+        glue.clear();
+    }
+
+    walk(
+        tree.root_node(),
+        source,
+        mod_prefix,
+        &[],
+        &mut units,
+        &mut glue,
+    );
+
+    // `<module>` is correct only when the file produced no real unit.
+    // Otherwise the import-only group becomes `<top-level>` (same
+    // post-pass as 1A Gap 1 / Python / TS).
+    let has_real_unit = units.iter().any(|(_, _, _, is_real)| *is_real);
+    if has_real_unit {
+        for (sym, _, _, is_real) in units.iter_mut() {
+            if !*is_real && sym.ends_with("<module>") {
+                let pre = &sym[..sym.len() - "<module>".len()];
+                *sym = format!("{pre}<top-level>");
+            }
+        }
+    }
+
+    let total_lines = lines.len() as u32;
+    let mut blocks = Vec::with_capacity(units.len());
+    for (ordinal, (symbol, ls, le, _is_real)) in units.into_iter().enumerate() {
+        let line_start = ls.max(1);
+        let line_end = le.min(total_lines.max(1));
+        let span = SourceSpan::Code {
+            line_start,
+            line_end,
+            symbol: Some(symbol),
+            lang: Some("javascript".to_string()),
+        };
+        let block_id = id_for_block(doc_id, "code", &[], ordinal as u32, &span);
+        let code = lines[(line_start as usize - 1)..=(line_end as usize - 1)].join("\n");
+        blocks.push(Block::Code(CodeBlock {
+            common: CommonBlock {
+                block_id,
+                heading_path: Vec::new(),
+                source_span: span,
+            },
+            lang: Some("javascript".to_string()),
+            code,
+        }));
+    }
+    Ok(blocks)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture(workspace_path: &str) -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(
+            concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/sample.js"),
+        )
+        .unwrap();
+        let asset = crate::rust::tests_support::fixed_code_asset(workspace_path, "javascript");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        JavascriptAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+    fn symbols(doc: &kebab_core::CanonicalDocument) -> Vec<String> {
+        let mut s: Vec<String> = doc
+            .blocks
+            .iter()
+            .filter_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, lang, .. } => {
+                        assert_eq!(lang.as_deref(), Some("javascript"));
+                        symbol.clone()
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .collect();
+        s.sort();
+        s
+    }
+    #[test]
+    fn extractor_supports_only_media_code_javascript() {
+        let e = JavascriptAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("javascript".into())));
+        assert!(!e.supports(&MediaType::Code("typescript".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+    #[test]
+    fn js_units_match_design_3_4_symbols() {
+        let doc = extract_fixture("src/sample.js");
+        let syms = symbols(&doc);
+        assert!(syms.iter().any(|s| s == "src/sample.add"), "got {syms:?}");
+        assert!(syms.iter().any(|s| s == "src/sample.Retriever"));
+        assert!(syms.iter().any(|s| s == "src/sample.Retriever.search"));
+        assert!(syms.iter().any(|s| s == "src/sample.Retriever.create"));
+        assert!(syms.iter().any(|s| s == "src/sample.default"));
+        assert!(syms.iter().any(|s| s == "src/sample.<top-level>"));
+    }
+    #[test]
+    fn jsx_via_js_grammar() {
+        // tree-sitter-javascript handles .jsx via the same single grammar.
+        let bytes = b"export function App() { return null; }\n";
+        let asset = crate::rust::tests_support::fixed_code_asset("src/App.jsx", "javascript");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        let doc = JavascriptAstExtractor::new().extract(&ctx, bytes).unwrap();
+        let syms = symbols(&doc);
+        assert!(syms.iter().any(|s| s == "src/App.App"), "got {syms:?}");
+    }
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture("src/sample.js");
+        for _ in 0..30 {
+            assert_eq!(extract_fixture("src/sample.js").blocks, a.blocks);
+        }
+    }
+
+    /// In tree-sitter-javascript, `decorator` is a CHILD of
+    /// `method_definition` (stored in the `decorator` field), so
+    /// `method_definition.start_row` already covers the decorator line
+    /// without any sibling walk. Verify that the emitted unit already
+    /// includes the decorator line and line_start is 2 (the @Log() line).
+    #[test]
+    fn js_class_method_decorator_already_folded_by_grammar() {
+        // Line 1 (1-indexed): "class Foo {"
+        // Line 2:             "    @Log()"   <- decorator (child of method_definition in JS grammar)
+        // Line 3:             "    bar() { return 1; }"
+        // Line 4:             "}"
+        let bytes = b"class Foo {\n    @Log()\n    bar() { return 1; }\n}\n";
+        let asset = crate::rust::tests_support::fixed_code_asset("src/foo.js", "javascript");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        let doc = JavascriptAstExtractor::new().extract(&ctx, bytes).unwrap();
+
+        let bar_block = doc
+            .blocks
+            .iter()
+            .find_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, .. }
+                        if symbol.as_deref() == Some("src/foo.Foo.bar") =>
+                    {
+                        Some(c)
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .expect("src/foo.Foo.bar block should be present");
+
+        // JS grammar: method_definition.start_row == decorator row, so
+        // no sibling walk change needed -- decorator is already included.
+        assert!(
+            bar_block.code.contains("@Log()"),
+            "JS method unit must include decorator (grammar folds it natively); got: {:?}",
+            bar_block.code
+        );
+        match &bar_block.common.source_span {
+            SourceSpan::Code { line_start, .. } => {
+                assert_eq!(
+                    *line_start, 2,
+                    "JS line_start must cover the @Log() decorator line (got {line_start})"
+                );
+            }
+            _ => unreachable!(),
+        }
+    }
+}

crates/kebab-parse-code/src/kotlin.rs

@@ -0,0 +1,627 @@
+//! `kebab-parse-code::kotlin` — tree-sitter Kotlin AST extractor (P10-1C-JK Task G).
+//!
+//! Implements [`kebab_core::Extractor`] for [`MediaType::Code("kotlin")`].
+//! Mirrors the Java extractor (JVM family, source-side `package` extraction +
+//! class-nesting) with Kotlin-specific adjustments:
+//!
+//! * Root is `source_file` (not `program`).
+//! * `package_header` carries a single `qualified_identifier` child whose
+//!   slice text IS the dotted package path — never a bare `identifier`
+//!   sub-form for the package (the grammar always wraps a single segment
+//!   in `qualified_identifier` too).
+//! * `class_declaration` covers `class`, `data class`, `sealed class`,
+//!   `enum class`, AND `interface` — Kotlin uses ONE node kind with a
+//!   `modifiers` child rather than separate `interface_declaration` /
+//!   `enum_declaration` nodes (verified via tree-sitter-kotlin-ng
+//!   `node-types.json`).
+//! * The body child of `class_declaration` is either `class_body` (normal
+//!   classes / interfaces) OR `enum_class_body` (enum class). Neither
+//!   carries a `body` field name, so it is matched by kind, not by
+//!   `child_by_field_name("body")`.
+//! * `companion_object` is a SEPARATE node kind (not `object_declaration`
+//!   with a modifier). Its `name` field is OPTIONAL — when omitted (the
+//!   common case `companion object { ... }`) the symbol uses the
+//!   implicit Kotlin convention name `Companion`.
+//! * `object_declaration` (named singleton) carries a `name` field and a
+//!   `class_body` child.
+//! * `function_declaration` may appear at top level (Kotlin top-level
+//!   function) AND inside `class_body` — same node kind, the
+//!   `mod_path` state distinguishes the two emit forms.
+//!
+//! Enum bodies (`enum_class_body`) are NOT recursed for the 1차 cut —
+//! `enum_entry` declarations are not emitted as units, matching the
+//! Java extractor's enum policy (design §3.4 1차 scope).
+//!
+//! Per design §3.4 / §9.1 / §9 versioning.
+
+use anyhow::Result;
+use kebab_core::{
+    Block, CanonicalDocument, CodeBlock, CommonBlock, Extractor, Lang, MediaType, Metadata,
+    ParserVersion, Provenance, ProvenanceEvent, ProvenanceKind, SourceSpan, SourceType, TrustLevel,
+    id_for_block, id_for_doc,
+};
+use serde_json::Map;
+use time::OffsetDateTime;
+
+use crate::scaffold::{filename_from_workspace_path, join_symbol, strip_extension};
+
+pub const PARSER_VERSION: &str = "code-kotlin-v1";
+
+/// Kotlin AST extractor. Per-unit blocks via tree-sitter-kotlin-ng 1.1
+/// (`LANGUAGE: LanguageFn`) parsed by tree-sitter 0.26.
+pub struct KotlinAstExtractor;
+
+impl KotlinAstExtractor {
+    pub fn new() -> Self {
+        Self
+    }
+}
+
+impl Default for KotlinAstExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl Extractor for KotlinAstExtractor {
+    fn supports(&self, m: &MediaType) -> bool {
+        matches!(m, MediaType::Code(l) if l == "kotlin")
+    }
+
+    fn parser_version(&self) -> ParserVersion {
+        ParserVersion(PARSER_VERSION.to_string())
+    }
+
+    fn extract(
+        &self,
+        ctx: &kebab_core::ExtractContext<'_>,
+        bytes: &[u8],
+    ) -> Result<CanonicalDocument> {
+        let asset = ctx.asset;
+        if !self.supports(&asset.media_type) {
+            anyhow::bail!(
+                "kebab-parse-code: unsupported media_type for KotlinAstExtractor: {:?}",
+                asset.media_type
+            );
+        }
+
+        let parser_version = self.parser_version();
+        let doc_id = id_for_doc(&asset.workspace_path, &asset.asset_id, &parser_version);
+
+        let source = String::from_utf8(bytes.to_vec()).map_err(|e| {
+            anyhow::anyhow!("kebab-parse-code: Kotlin source is not valid UTF-8: {e}")
+        })?;
+
+        let blocks = build_blocks(&source, &doc_id)?;
+        let unit_count = blocks.len() as u32;
+
+        let now = OffsetDateTime::now_utc();
+        let mut events: Vec<ProvenanceEvent> = Vec::with_capacity(2);
+        events.push(ProvenanceEvent {
+            at: asset.discovered_at,
+            agent: "kb-source-fs".to_string(),
+            kind: ProvenanceKind::Discovered,
+            note: None,
+        });
+        events.push(ProvenanceEvent {
+            at: now,
+            agent: "kb-parse-code".to_string(),
+            kind: ProvenanceKind::Parsed,
+            note: Some(format!(
+                "parser_version={}; unit_count={}",
+                parser_version.0, unit_count
+            )),
+        });
+
+        let title = {
+            let fname = filename_from_workspace_path(&asset.workspace_path.0);
+            strip_extension(&fname)
+        };
+
+        // Resolve the file's absolute path for repo detection. If the
+        // source URI carries a relative path, anchor it at the workspace
+        // root so the `.git/` walk-up starts from the right place.
+        let abs_path = match &asset.source_uri {
+            kebab_core::SourceUri::File(p) => {
+                if p.is_absolute() {
+                    p.clone()
+                } else {
+                    ctx.workspace_root.join(p)
+                }
+            }
+            kebab_core::SourceUri::Kb(_) => ctx.workspace_root.to_path_buf(),
+        };
+        let (repo, git_branch, git_commit) = match crate::repo::detect_repo(&abs_path) {
+            Some(r) => (Some(r.name), r.branch, r.commit),
+            None => (None, None, None),
+        };
+
+        let metadata = Metadata {
+            aliases: Vec::new(),
+            tags: Vec::new(),
+            created_at: asset.discovered_at,
+            updated_at: asset.discovered_at,
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Map::new(),
+            repo,
+            git_branch,
+            git_commit,
+            code_lang: Some("kotlin".to_string()),
+        };
+
+        tracing::debug!(
+            target: "kebab-parse-code",
+            "extracted Kotlin doc_id={} workspace_path={} units={}",
+            doc_id.0,
+            asset.workspace_path.0,
+            unit_count
+        );
+
+        Ok(CanonicalDocument {
+            doc_id,
+            source_asset_id: asset.asset_id.clone(),
+            workspace_path: asset.workspace_path.clone(),
+            title,
+            lang: Lang("und".to_string()),
+            blocks,
+            metadata,
+            provenance: Provenance { events },
+            parser_version,
+            schema_version: 1,
+            doc_version: 1,
+            last_chunker_version: None,
+            last_embedding_version: None,
+        })
+    }
+}
+
+/// p10-1C-JK: extract `package` declaration text from a tree-sitter-kotlin
+/// `source_file`. Returns `None` if no `package_header` (default-package
+/// Kotlin file). The package_header's single named child is a
+/// `qualified_identifier`; its slice text is the dotted path. Per design
+/// §3.4 Kotlin row.
+fn extract_package(root: tree_sitter::Node, src: &str) -> Option<String> {
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        if child.kind() == "package_header" {
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                let k = sub.kind();
+                if k == "qualified_identifier" || k == "identifier" {
+                    return Some(src[sub.start_byte()..sub.end_byte()].to_string());
+                }
+            }
+        }
+    }
+    None
+}
+
+/// Walk preceding `line_comment` / `block_comment` siblings to extend
+/// the unit's line range upward, folding leading KDoc / line comments
+/// into the unit. Modifiers / annotations live INSIDE the declaration
+/// node itself, so their lines are already inside `n.start_position()`.
+fn unit_start(n: &tree_sitter::Node) -> u32 {
+    let mut start = n.start_position().row as u32 + 1;
+    let mut prev = n.prev_sibling();
+    while let Some(p) = prev {
+        let k = p.kind();
+        if k == "line_comment" || k == "block_comment" {
+            start = p.start_position().row as u32 + 1;
+            prev = p.prev_sibling();
+        } else {
+            break;
+        }
+    }
+    start
+}
+
+fn node_name_text<'a>(n: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+    n.child_by_field_name("name")
+        .map(|c| &src[c.start_byte()..c.end_byte()])
+}
+
+/// Find the first child of a node with one of the given kinds. Used to
+/// locate `class_body` / `enum_class_body` on `class_declaration` since
+/// the kotlin grammar attaches them without a `body` field name.
+fn first_child_of_kinds<'a>(
+    n: &tree_sitter::Node<'a>,
+    kinds: &[&str],
+) -> Option<tree_sitter::Node<'a>> {
+    let mut cur = n.walk();
+    n.named_children(&mut cur)
+        .find(|child| kinds.contains(&child.kind()))
+}
+
+/// `true` iff a `class_declaration` carries the `enum` class modifier.
+/// Detected by walking `modifiers` → `class_modifier` and checking the
+/// child text. The grammar exposes "enum" / "sealed" / "data" /
+/// "annotation" / "inner" as named `class_modifier` children of
+/// `modifiers`. We only need to know about "enum" to decide whether to
+/// look for `class_body` or `enum_class_body` and whether to skip body
+/// recursion.
+fn class_decl_is_enum(n: &tree_sitter::Node, src: &str) -> bool {
+    let mut cur = n.walk();
+    for child in n.named_children(&mut cur) {
+        if child.kind() == "modifiers" {
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                if sub.kind() == "class_modifier" {
+                    let text = &src[sub.start_byte()..sub.end_byte()];
+                    if text == "enum" {
+                        return true;
+                    }
+                }
+            }
+        }
+    }
+    false
+}
+
+fn build_blocks(
+    source: &str,
+    doc_id: &kebab_core::DocumentId,
+) -> anyhow::Result<Vec<kebab_core::Block>> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&tree_sitter_kotlin_ng::LANGUAGE.into())
+        .map_err(|e| anyhow::anyhow!("set tree-sitter-kotlin-ng language: {e}"))?;
+    let tree = parser
+        .parse(source.as_bytes(), None)
+        .ok_or_else(|| anyhow::anyhow!("tree-sitter failed to parse Kotlin source"))?;
+    let lines: Vec<&str> = source.split('\n').collect();
+
+    let root = tree.root_node();
+    let mod_prefix = extract_package(root, source).unwrap_or_else(|| "<unknown>".to_string());
+
+    // units: (symbol, line_start, line_end, is_real_semantic_unit).
+    // Glue groups are pushed with a sentinel symbol + is_real=false so a
+    // post-pass can decide `<module>` vs `<top-level>` (JVM family pattern).
+    let mut units: Vec<(String, u32, u32, bool)> = Vec::new();
+    // (is_import 0/1, s, e). `is_import` flags `import` — used by the
+    // glue flush to pick `<module>` vs `<top-level>` provisional label.
+    let mut glue: Vec<(usize, u32, u32)> = Vec::new();
+
+    walk_top(root, source, &mod_prefix, &mut units, &mut glue);
+
+    // `<module>` is correct only when the file produced no real unit.
+    // Otherwise the import-only group becomes `<top-level>` (same
+    // post-pass as 1B / 1C-Go / Java).
+    let has_real_unit = units.iter().any(|(_, _, _, is_real)| *is_real);
+    if has_real_unit {
+        for (sym, _, _, is_real) in units.iter_mut() {
+            if !*is_real && sym.ends_with("<module>") {
+                let pre = &sym[..sym.len() - "<module>".len()];
+                *sym = format!("{pre}<top-level>");
+            }
+        }
+    }
+
+    let total_lines = lines.len() as u32;
+    let mut blocks = Vec::with_capacity(units.len());
+    for (ordinal, (symbol, ls, le, _is_real)) in units.into_iter().enumerate() {
+        let line_start = ls.max(1);
+        let line_end = le.min(total_lines.max(1));
+        let span = SourceSpan::Code {
+            line_start,
+            line_end,
+            symbol: Some(symbol),
+            lang: Some("kotlin".to_string()),
+        };
+        let block_id = id_for_block(doc_id, "code", &[], ordinal as u32, &span);
+        let code = lines[(line_start as usize - 1)..=(line_end as usize - 1)].join("\n");
+        blocks.push(Block::Code(CodeBlock {
+            common: CommonBlock {
+                block_id,
+                heading_path: Vec::new(),
+                source_span: span,
+            },
+            lang: Some("kotlin".to_string()),
+            code,
+        }));
+    }
+    Ok(blocks)
+}
+
+/// Walk the file's top-level children — `source_file` named children:
+/// `package_header` (handled by `extract_package`), `import` (glue),
+/// `class_declaration` (class / interface / enum class), `object_declaration`,
+/// `function_declaration` (top-level), `property_declaration` (top-level),
+/// `type_alias` (currently treated as glue). Class / object bodies are
+/// recursed via [`walk_body`] with the type name pushed onto `mod_path`
+/// (JVM family pattern). Enum bodies are NOT recursed (1차 cut).
+fn walk_top(
+    node: tree_sitter::Node,
+    src: &str,
+    mod_prefix: &str,
+    units: &mut Vec<(String, u32, u32, bool)>,
+    glue: &mut Vec<(usize, u32, u32)>,
+) {
+    let mod_path: &[String] = &[];
+    let mut cur = node.walk();
+    for child in node.named_children(&mut cur) {
+        let s = unit_start(&child);
+        let e = child.end_position().row as u32 + 1;
+        match child.kind() {
+            "class_declaration" => {
+                // Covers class / data class / sealed class / interface /
+                // enum class — single grammar node, the modifiers child
+                // distinguishes them. The body is `class_body` for
+                // non-enum and `enum_class_body` for enum class; both
+                // attach without a `body` field name.
+                if let Some(name) = node_name_text(&child, src) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(glue, units, mod_prefix, mod_path);
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                    let is_enum = class_decl_is_enum(&child, src);
+                    if !is_enum {
+                        if let Some(body) = first_child_of_kinds(&child, &["class_body"]) {
+                            let np: Vec<String> = vec![name.to_string()];
+                            walk_body(body, src, mod_prefix, &np, units);
+                        }
+                    }
+                    // enum_class_body NOT recursed — enum constants are
+                    // not emitted as units (1차 scope, matches Java).
+                }
+            }
+            "object_declaration" => {
+                // Singleton object — name field is required by the grammar.
+                if let Some(name) = node_name_text(&child, src) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(glue, units, mod_prefix, mod_path);
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                    if let Some(body) = first_child_of_kinds(&child, &["class_body"]) {
+                        let np: Vec<String> = vec![name.to_string()];
+                        walk_body(body, src, mod_prefix, &np, units);
+                    }
+                }
+            }
+            "function_declaration" => {
+                // Top-level Kotlin function (unlike Java).
+                if let Some(name) = node_name_text(&child, src) {
+                    glue.retain(|(_, gs, _)| *gs < s);
+                    flush_glue(glue, units, mod_prefix, mod_path);
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                }
+            }
+            "import" => {
+                glue.push((1, s, e));
+            }
+            // `property_declaration` (top-level val/var) and `type_alias`
+            // are not emitted as standalone units in the 1차 cut — they
+            // glue into the import group instead. `package_header` is
+            // handled by `extract_package` (structural metadata, not a
+            // unit).
+            _ => {}
+        }
+    }
+    flush_glue(glue, units, mod_prefix, mod_path);
+}
+
+/// Walk a `class_body` (or object's `class_body`). Emits one unit per
+/// method / secondary constructor and recurses into nested type
+/// declarations + companion objects. Property declarations are NOT
+/// emitted (would explode unit count, parallel to Java field policy).
+///
+/// `companion_object` carries an optional `name` field — when omitted
+/// (the common case `companion object { ... }`) the implicit Kotlin
+/// convention name `Companion` is used.
+///
+/// No `glue` parameter: Kotlin imports are file-level only.
+fn walk_body(
+    body: tree_sitter::Node,
+    src: &str,
+    mod_prefix: &str,
+    mod_path: &[String],
+    units: &mut Vec<(String, u32, u32, bool)>,
+) {
+    let mut cur = body.walk();
+    for child in body.named_children(&mut cur) {
+        let s = unit_start(&child);
+        let e = child.end_position().row as u32 + 1;
+        match child.kind() {
+            "function_declaration" => {
+                if let Some(name) = node_name_text(&child, src) {
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                }
+            }
+            "secondary_constructor" => {
+                // Kotlin secondary constructor — no `name` field on the
+                // grammar node. Per design §3.4 (Java JVM convention) the
+                // symbol uses the enclosing class name as the trailing
+                // segment (matches the Java `<pkg>.<...>.<Class>.<Class>`
+                // duplication for constructors).
+                if let Some(class_name) = mod_path.last() {
+                    let sym = join_symbol(mod_prefix, mod_path, class_name);
+                    units.push((sym, s, e, true));
+                }
+            }
+            "companion_object" => {
+                // Companion's name field is OPTIONAL — fall back to the
+                // Kotlin implicit name `Companion`.
+                let name: &str = node_name_text(&child, src).unwrap_or("Companion");
+                let sym = join_symbol(mod_prefix, mod_path, name);
+                units.push((sym, s, e, true));
+                if let Some(inner_body) = first_child_of_kinds(&child, &["class_body"]) {
+                    let mut np = mod_path.to_vec();
+                    np.push(name.to_string());
+                    walk_body(inner_body, src, mod_prefix, &np, units);
+                }
+            }
+            "class_declaration" => {
+                let name = match node_name_text(&child, src) {
+                    Some(n) => n,
+                    None => continue,
+                };
+                let sym = join_symbol(mod_prefix, mod_path, name);
+                units.push((sym, s, e, true));
+                let is_enum = class_decl_is_enum(&child, src);
+                if !is_enum {
+                    if let Some(inner_body) = first_child_of_kinds(&child, &["class_body"]) {
+                        let mut np = mod_path.to_vec();
+                        np.push(name.to_string());
+                        walk_body(inner_body, src, mod_prefix, &np, units);
+                    }
+                }
+            }
+            "object_declaration" => {
+                let name = match node_name_text(&child, src) {
+                    Some(n) => n,
+                    None => continue,
+                };
+                let sym = join_symbol(mod_prefix, mod_path, name);
+                units.push((sym, s, e, true));
+                if let Some(inner_body) = first_child_of_kinds(&child, &["class_body"]) {
+                    let mut np = mod_path.to_vec();
+                    np.push(name.to_string());
+                    walk_body(inner_body, src, mod_prefix, &np, units);
+                }
+            }
+            // property_declaration, anonymous_initializer: NOT emitted.
+            _ => {}
+        }
+    }
+}
+
+fn flush_glue(
+    glue: &mut Vec<(usize, u32, u32)>,
+    units: &mut Vec<(String, u32, u32, bool)>,
+    mod_prefix: &str,
+    mod_path: &[String],
+) {
+    if glue.is_empty() {
+        return;
+    }
+    let s = glue.iter().map(|(_, a, _)| *a).min().unwrap();
+    let e = glue.iter().map(|(_, _, b)| *b).max().unwrap();
+    // Provisional label: `<module>` only if the group is exclusively
+    // imports. The post-pass demotes any `<module>` to `<top-level>` if
+    // the file produced any real unit.
+    let only_imports = glue.iter().all(|(is_import, _, _)| *is_import == 1);
+    let label = if only_imports { "<module>" } else { "<top-level>" };
+    units.push((join_symbol(mod_prefix, mod_path, label), s, e, false));
+    glue.clear();
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(concat!(
+            env!("CARGO_MANIFEST_DIR"),
+            "/tests/fixtures/sample.kt"
+        ))
+        .unwrap();
+        let asset =
+            crate::rust::tests_support::fixed_code_asset("crates/x/src/sample.kt", "kotlin");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        KotlinAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_kotlin() {
+        let e = KotlinAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("kotlin".into())));
+        assert!(!e.supports(&MediaType::Code("java".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+
+    #[test]
+    fn kotlin_units_match_design_3_4_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc
+            .blocks
+            .iter()
+            .filter_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, lang, .. } => {
+                        assert_eq!(lang.as_deref(), Some("kotlin"));
+                        symbol.clone()
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .collect();
+        syms.sort();
+        // package extracted from source = com.kebab.chunk
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.chunkDoc"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.getName"),
+            "got {syms:?}"
+        );
+        // Implicit companion object name = Companion (grammar leaves the
+        // name field unset; the extractor fills it in).
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Companion"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter()
+                .any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Companion.withName"),
+            "got {syms:?}"
+        );
+        // interface — also via class_declaration in the grammar
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.Stringer"),
+            "got {syms:?}"
+        );
+        // enum class — also via class_declaration; body NOT recursed
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.Mode"),
+            "got {syms:?}"
+        );
+        // Kotlin top-level fn — unlike Java
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.freeFunction"),
+            "got {syms:?}"
+        );
+        // Singleton object + its method
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.Singleton"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.Singleton.ping"),
+            "got {syms:?}"
+        );
+        // import grouped as <top-level>
+        assert!(
+            syms.iter().any(|s| s == "com.kebab.chunk.<top-level>"),
+            "got {syms:?}"
+        );
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 {
+            assert_eq!(extract_fixture().blocks, a.blocks);
+        }
+    }
+}

crates/kebab-parse-code/src/lang.rs

@@ -24,7 +24,7 @@ pub fn code_lang_for_path(path: &Path) -> Option<&'static str> {
     match ext.as_str() {
         "rs" => Some("rust"),
         "py" | "pyi" => Some("python"),
-        "ts" | "tsx" => Some("typescript"),
+        "ts" | "tsx" | "mts" | "cts" => Some("typescript"),
         "js" | "mjs" | "cjs" | "jsx" => Some("javascript"),
         "go" => Some("go"),
         "java" => Some("java"),
@@ -40,3 +40,82 @@ pub fn code_lang_for_path(path: &Path) -> Option<&'static str> {
         _ => None,
     }
 }
+
+/// p10-1B: workspace-relative Python file path → dotted module-path prefix.
+/// See plan §Task C for the exact rules + tasks/p10/p10-1b for the §3.4
+/// design contract.
+///
+/// Stripped source-roots: `src/`, `lib/`, and `crates/<crate>/src/`.
+/// `tests/`, `examples/`, and `benches/` are intentionally NOT stripped —
+/// they appear in test/example/bench namespaces and dropping them would
+/// conflate identical symbol names across conventional Python directories
+/// (e.g. `tests/test_foo.py` → `tests.test_foo`, not `test_foo`).
+pub fn module_path_for_python(workspace_path: &str) -> String {
+    let mut p: &str = workspace_path;
+    if let Some(rest) = p.strip_prefix("crates/") {
+        if let Some(slash) = rest.find('/') {
+            let after = &rest[slash + 1..];
+            if let Some(stripped) = after.strip_prefix("src/") {
+                p = stripped;
+            }
+        }
+    } else if let Some(stripped) = p.strip_prefix("src/") {
+        p = stripped;
+    } else if let Some(stripped) = p.strip_prefix("lib/") {
+        p = stripped;
+    }
+    let p = match p.strip_suffix(".py") {
+        Some(s) => s,
+        None => p.strip_suffix(".pyi").unwrap_or(p),
+    };
+    let p = if let Some(parent) = p.strip_suffix("/__init__") {
+        parent
+    } else if p == "__init__" {
+        ""
+    } else {
+        p
+    };
+    p.replace('/', ".")
+}
+
+/// p10-1B: workspace-relative TS/JS file path → path-style prefix
+/// (no slash replacement, no source-root strip). See plan §Task C.
+pub fn module_path_for_tsjs(workspace_path: &str) -> String {
+    let p = workspace_path;
+    for ext in [".tsx", ".mts", ".cts", ".ts", ".jsx", ".mjs", ".cjs", ".js"] {
+        if let Some(stripped) = p.strip_suffix(ext) {
+            return stripped.to_string();
+        }
+    }
+    p.to_string()
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn module_path_for_python_strips_src_roots_and_extensions() {
+        assert_eq!(module_path_for_python("kebab_eval/metrics.py"),       "kebab_eval.metrics");
+        assert_eq!(module_path_for_python("kebab_eval/__init__.py"),      "kebab_eval");
+        assert_eq!(module_path_for_python("src/foo/bar.py"),              "foo.bar");
+        assert_eq!(module_path_for_python("crates/x/src/foo/bar.py"),     "foo.bar");
+        assert_eq!(module_path_for_python("a/b/c.pyi"),                   "a.b.c");
+        assert_eq!(module_path_for_python("standalone.py"),               "standalone");
+        assert_eq!(module_path_for_python("src/__init__.py"),             "");
+        // `tests/` is NOT a stripped source-root — it is preserved as
+        // part of the module path so test symbols stay namespaced.
+        assert_eq!(module_path_for_python("tests/test_foo.py"),           "tests.test_foo");
+    }
+
+    #[test]
+    fn module_path_for_tsjs_keeps_slashes_and_strips_ext() {
+        for ext in ["ts", "tsx", "mts", "cts", "js", "jsx", "mjs", "cjs"] {
+            let p = format!("src/search/retriever/Retriever.{ext}");
+            assert_eq!(module_path_for_tsjs(&p), "src/search/retriever/Retriever");
+        }
+        assert_eq!(module_path_for_tsjs("foo.ts"),                 "foo");
+        assert_eq!(module_path_for_tsjs("a/b/c.ts"),               "a/b/c");
+        assert_eq!(module_path_for_tsjs("packages/x/src/Foo.ts"),  "packages/x/src/Foo");
+    }
+}

crates/kebab-parse-code/src/lib.rs

@@ -13,12 +13,25 @@
 //! `kebab-parse-*` crates per design §8: must NOT depend on store / embed
 //! / llm / rag.
 
+pub mod go;
+pub mod java;
+pub mod javascript;
+pub mod kotlin;
 pub mod lang;
+pub mod python;
 pub mod repo;
 pub mod rust;
+pub(crate) mod scaffold;
 pub mod skip;
+pub mod typescript;
 
-pub use lang::code_lang_for_path;
+pub use go::{PARSER_VERSION as GO_PARSER_VERSION, GoAstExtractor};
+pub use java::{PARSER_VERSION as JAVA_PARSER_VERSION, JavaAstExtractor};
+pub use javascript::{PARSER_VERSION as JS_PARSER_VERSION, JavascriptAstExtractor};
+pub use kotlin::{PARSER_VERSION as KOTLIN_PARSER_VERSION, KotlinAstExtractor};
+pub use lang::{code_lang_for_path, module_path_for_python, module_path_for_tsjs};
+pub use python::{PARSER_VERSION as PYTHON_PARSER_VERSION, PythonAstExtractor};
 pub use repo::{RepoMeta, detect_repo};
 pub use rust::{PARSER_VERSION as RUST_PARSER_VERSION, RustAstExtractor};
 pub use skip::{BUILTIN_BLACKLIST, is_generated_file, is_oversized};
+pub use typescript::{PARSER_VERSION as TS_PARSER_VERSION, TypescriptAstExtractor};

crates/kebab-parse-code/src/python.rs

@@ -0,0 +1,437 @@
+//! `kebab-parse-code::python` — tree-sitter Python AST extractor (P10-1B Task E).
+//!
+//! Implements [`kebab_core::Extractor`] for [`MediaType::Code("python")`].
+//! Walks the tree-sitter parse tree and emits one [`Block::Code`] per
+//! top-level AST semantic unit (free fn, class, each method, recursively
+//! per nested class), each carrying [`SourceSpan::Code`] with the unit's
+//! dotted self-reference symbol path prefixed by `module_path_for_python`
+//! (design §3.4). Glue declarations (`import` / `import from` /
+//! `expression_statement` / `assignment` / `global_statement` /
+//! `future_import_statement`) collapse into one grouped `<top-level>`
+//! (or `<module>`) unit.
+//!
+//! Decorators are folded into the decorated unit's line range via the
+//! `decorated_definition` unwrap arm (analog of the Rust `attribute_item`
+//! re-absorption in 1A — see §9.1).
+//!
+//! Scope follows 1A: AST unit extraction + dotted symbol paths + line
+//! ranges. Per design §3.4 / §9.1 / §9 versioning.
+
+use anyhow::Result;
+use kebab_core::{
+    Block, CanonicalDocument, CodeBlock, CommonBlock, Extractor, Lang, MediaType, Metadata,
+    ParserVersion, Provenance, ProvenanceEvent, ProvenanceKind, SourceSpan, SourceType, TrustLevel,
+    id_for_block, id_for_doc,
+};
+use serde_json::Map;
+use time::OffsetDateTime;
+
+use crate::scaffold::{filename_from_workspace_path, join_symbol, strip_extension};
+
+pub const PARSER_VERSION: &str = "code-python-v1";
+
+/// Python AST extractor. Per-unit blocks via tree-sitter-python 0.25
+/// (`LANGUAGE: LanguageFn`) parsed by tree-sitter 0.26.
+pub struct PythonAstExtractor;
+
+impl PythonAstExtractor {
+    pub fn new() -> Self {
+        Self
+    }
+}
+
+impl Default for PythonAstExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl Extractor for PythonAstExtractor {
+    fn supports(&self, m: &MediaType) -> bool {
+        matches!(m, MediaType::Code(l) if l == "python")
+    }
+
+    fn parser_version(&self) -> ParserVersion {
+        ParserVersion(PARSER_VERSION.to_string())
+    }
+
+    fn extract(
+        &self,
+        ctx: &kebab_core::ExtractContext<'_>,
+        bytes: &[u8],
+    ) -> Result<CanonicalDocument> {
+        let asset = ctx.asset;
+        if !self.supports(&asset.media_type) {
+            anyhow::bail!(
+                "kebab-parse-code: unsupported media_type for PythonAstExtractor: {:?}",
+                asset.media_type
+            );
+        }
+
+        let parser_version = self.parser_version();
+        let doc_id = id_for_doc(&asset.workspace_path, &asset.asset_id, &parser_version);
+
+        let source = String::from_utf8(bytes.to_vec()).map_err(|e| {
+            anyhow::anyhow!("kebab-parse-code: Python source is not valid UTF-8: {e}")
+        })?;
+
+        let mod_prefix = crate::lang::module_path_for_python(&asset.workspace_path.0);
+        let blocks = build_blocks(&source, &doc_id, &mod_prefix)?;
+        let unit_count = blocks.len() as u32;
+
+        let now = OffsetDateTime::now_utc();
+        let mut events: Vec<ProvenanceEvent> = Vec::with_capacity(2);
+        events.push(ProvenanceEvent {
+            at: asset.discovered_at,
+            agent: "kb-source-fs".to_string(),
+            kind: ProvenanceKind::Discovered,
+            note: None,
+        });
+        events.push(ProvenanceEvent {
+            at: now,
+            agent: "kb-parse-code".to_string(),
+            kind: ProvenanceKind::Parsed,
+            note: Some(format!(
+                "parser_version={}; unit_count={}",
+                parser_version.0, unit_count
+            )),
+        });
+
+        let title = {
+            let fname = filename_from_workspace_path(&asset.workspace_path.0);
+            strip_extension(&fname)
+        };
+
+        // Resolve the file's absolute path for repo detection. If the
+        // source URI carries a relative path, anchor it at the workspace
+        // root so the `.git/` walk-up starts from the right place.
+        let abs_path = match &asset.source_uri {
+            kebab_core::SourceUri::File(p) => {
+                if p.is_absolute() {
+                    p.clone()
+                } else {
+                    ctx.workspace_root.join(p)
+                }
+            }
+            kebab_core::SourceUri::Kb(_) => ctx.workspace_root.to_path_buf(),
+        };
+        let (repo, git_branch, git_commit) = match crate::repo::detect_repo(&abs_path) {
+            Some(r) => (Some(r.name), r.branch, r.commit),
+            None => (None, None, None),
+        };
+
+        let metadata = Metadata {
+            aliases: Vec::new(),
+            tags: Vec::new(),
+            created_at: asset.discovered_at,
+            updated_at: asset.discovered_at,
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Map::new(),
+            repo,
+            git_branch,
+            git_commit,
+            code_lang: Some("python".to_string()),
+        };
+
+        tracing::debug!(
+            target: "kebab-parse-code",
+            "extracted Python doc_id={} workspace_path={} units={}",
+            doc_id.0,
+            asset.workspace_path.0,
+            unit_count
+        );
+
+        Ok(CanonicalDocument {
+            doc_id,
+            source_asset_id: asset.asset_id.clone(),
+            workspace_path: asset.workspace_path.clone(),
+            title,
+            lang: Lang("und".to_string()),
+            blocks,
+            metadata,
+            provenance: Provenance { events },
+            parser_version,
+            schema_version: 1,
+            doc_version: 1,
+            last_chunker_version: None,
+            last_embedding_version: None,
+        })
+    }
+}
+
+fn build_blocks(
+    source: &str,
+    doc_id: &kebab_core::DocumentId,
+    mod_prefix: &str,
+) -> anyhow::Result<Vec<kebab_core::Block>> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&tree_sitter_python::LANGUAGE.into())
+        .map_err(|e| anyhow::anyhow!("set tree-sitter-python language: {e}"))?;
+    let tree = parser
+        .parse(source.as_bytes(), None)
+        .ok_or_else(|| anyhow::anyhow!("tree-sitter failed to parse Python source"))?;
+    let lines: Vec<&str> = source.split('\n').collect();
+
+    // units: (symbol, line_start, line_end, is_real_semantic_unit).
+    // Glue groups are pushed with a sentinel symbol + is_real=false so a
+    // post-pass can decide `<module>` vs `<top-level>` (same algorithm
+    // as 1A Gap 1).
+    let mut units: Vec<(String, u32, u32, bool)> = Vec::new();
+    // (is_import 0/1, s, e). `is_import` flags `import_statement` /
+    // `import_from_statement` / `future_import_statement` — used by the
+    // glue flush to pick `<module>` vs `<top-level>` provisional label
+    // (1A's `is_mod_decl` analog).
+    let mut glue: Vec<(usize, u32, u32)> = Vec::new();
+
+    fn node_name<'a>(n: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+        n.child_by_field_name("name")
+            .map(|c| &src[c.start_byte()..c.end_byte()])
+    }
+    /// Walk preceding `comment` siblings to extend the unit's line range
+    /// upward, folding leading doc / line comments into the unit. Note
+    /// that Python decorators are NOT preceding siblings — they live
+    /// INSIDE a `decorated_definition` parent — so they are handled by
+    /// the unwrap arm below, not here.
+    fn unit_start(n: &tree_sitter::Node) -> u32 {
+        let mut start = n.start_position().row as u32 + 1;
+        let mut prev = n.prev_sibling();
+        while let Some(p) = prev {
+            if p.kind() == "comment" {
+                start = p.start_position().row as u32 + 1;
+                prev = p.prev_sibling();
+            } else {
+                break;
+            }
+        }
+        start
+    }
+    fn walk(
+        node: tree_sitter::Node,
+        src: &str,
+        mod_prefix: &str,
+        mod_path: &[String],
+        units: &mut Vec<(String, u32, u32, bool)>,
+        glue: &mut Vec<(usize, u32, u32)>,
+    ) {
+        let mut cur = node.walk();
+        for child in node.named_children(&mut cur) {
+            // Default unit line range — overridden by the
+            // `decorated_definition` unwrap arm so decorator lines are
+            // included.
+            let s = unit_start(&child);
+            let e = child.end_position().row as u32 + 1;
+            match child.kind() {
+                "function_definition" => {
+                    if let Some(name) = node_name(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                    }
+                }
+                "class_definition" => {
+                    if let Some(name) = node_name(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                        // Recurse into the class body with the class
+                        // name pushed onto mod_path; methods become
+                        // `<...>.<ClassName>.<method>` and nested
+                        // classes recurse further with both names.
+                        if let Some(body) = child.child_by_field_name("body") {
+                            let mut np = mod_path.to_vec();
+                            np.push(name.to_string());
+                            walk(body, src, mod_prefix, &np, units, glue);
+                            debug_assert!(
+                                glue.is_empty(),
+                                "inner walk must flush its glue before returning"
+                            );
+                        }
+                    }
+                }
+                "decorated_definition" => {
+                    // Unwrap: the inner definition supplies the symbol
+                    // name, but the unit's line range comes from the
+                    // OUTER `decorated_definition` so decorator lines
+                    // are folded in (analog of `attribute_item`
+                    // re-absorption in 1A — see plan §Task E note (b)).
+                    if let Some(inner) = child.child_by_field_name("definition") {
+                        let outer_s = s; // already includes decorators
+                        let outer_e = e;
+                        match inner.kind() {
+                            "function_definition" => {
+                                if let Some(name) = node_name(&inner, src) {
+                                    glue.retain(|(_, gs, _)| *gs < outer_s);
+                                    flush_glue(glue, units, mod_prefix, mod_path);
+                                    let sym = join_symbol(mod_prefix, mod_path, name);
+                                    units.push((sym, outer_s, outer_e, true));
+                                }
+                            }
+                            "class_definition" => {
+                                if let Some(name) = node_name(&inner, src) {
+                                    glue.retain(|(_, gs, _)| *gs < outer_s);
+                                    flush_glue(glue, units, mod_prefix, mod_path);
+                                    let sym = join_symbol(mod_prefix, mod_path, name);
+                                    units.push((sym, outer_s, outer_e, true));
+                                    if let Some(body) = inner.child_by_field_name("body") {
+                                        let mut np = mod_path.to_vec();
+                                        np.push(name.to_string());
+                                        walk(body, src, mod_prefix, &np, units, glue);
+                                        debug_assert!(
+                                            glue.is_empty(),
+                                            "inner walk must flush its glue before returning"
+                                        );
+                                    }
+                                }
+                            }
+                            _ => {}
+                        }
+                    }
+                }
+                "import_statement" | "import_from_statement" | "future_import_statement" => {
+                    glue.push((1, s, e));
+                }
+                "expression_statement" | "assignment" | "global_statement" => {
+                    glue.push((0, s, e));
+                }
+                _ => {}
+            }
+        }
+        flush_glue(glue, units, mod_prefix, mod_path);
+    }
+    fn flush_glue(
+        glue: &mut Vec<(usize, u32, u32)>,
+        units: &mut Vec<(String, u32, u32, bool)>,
+        mod_prefix: &str,
+        mod_path: &[String],
+    ) {
+        if glue.is_empty() {
+            return;
+        }
+        let s = glue.iter().map(|(_, a, _)| *a).min().unwrap();
+        let e = glue.iter().map(|(_, _, b)| *b).max().unwrap();
+        // Provisional label: `<module>` only if the group is exclusively
+        // imports (1A's `only_mod_decls` analog). The post-pass below
+        // demotes any `<module>` to `<top-level>` if the file produced
+        // any real unit.
+        let only_imports = glue.iter().all(|(is_import, _, _)| *is_import == 1);
+        let label = if only_imports { "<module>" } else { "<top-level>" };
+        units.push((join_symbol(mod_prefix, mod_path, label), s, e, false));
+        glue.clear();
+    }
+
+    walk(tree.root_node(), source, mod_prefix, &[], &mut units, &mut glue);
+
+    // `<module>` is correct only when the file produced no real unit.
+    // Otherwise the import-only group becomes `<top-level>` (same
+    // algorithm as 1A Gap 1). Match on the suffix so a class-nested
+    // glue group (which doesn't exist in current Python AST but is
+    // future-proofed) still demotes correctly.
+    let has_real_unit = units.iter().any(|(_, _, _, is_real)| *is_real);
+    if has_real_unit {
+        for (sym, _, _, is_real) in units.iter_mut() {
+            if !*is_real && sym.ends_with("<module>") {
+                let pre = &sym[..sym.len() - "<module>".len()];
+                *sym = format!("{pre}<top-level>");
+            }
+        }
+    }
+
+    let total_lines = lines.len() as u32;
+    let mut blocks = Vec::with_capacity(units.len());
+    for (ordinal, (symbol, ls, le, _is_real)) in units.into_iter().enumerate() {
+        let line_start = ls.max(1);
+        let line_end = le.min(total_lines.max(1));
+        let span = SourceSpan::Code {
+            line_start,
+            line_end,
+            symbol: Some(symbol),
+            lang: Some("python".to_string()),
+        };
+        let block_id = id_for_block(doc_id, "code", &[], ordinal as u32, &span);
+        let code = lines[(line_start as usize - 1)..=(line_end as usize - 1)].join("\n");
+        blocks.push(Block::Code(CodeBlock {
+            common: CommonBlock {
+                block_id,
+                heading_path: Vec::new(),
+                source_span: span,
+            },
+            lang: Some("python".to_string()),
+            code,
+        }));
+    }
+    Ok(blocks)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(
+            concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/sample.py"),
+        )
+        .unwrap();
+        let asset = crate::rust::tests_support::fixed_code_asset(
+            "kebab_eval/metrics.py", "python",
+        );
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset, workspace_root: &root, config: &cfg,
+        };
+        PythonAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_python() {
+        let e = PythonAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("python".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+
+    #[test]
+    fn python_units_carry_module_prefixed_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc.blocks.iter().map(|b| match b {
+            Block::Code(c) => match &c.common.source_span {
+                SourceSpan::Code { symbol, lang, .. } => {
+                    assert_eq!(lang.as_deref(), Some("python"));
+                    symbol.clone().unwrap()
+                }
+                _ => panic!("expected SourceSpan::Code"),
+            },
+            other => panic!("expected Block::Code, got {other:?}"),
+        }).collect();
+        syms.sort();
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.free"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Foo"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Foo.double"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Foo.name"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Outer"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Outer.Inner"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Outer.Inner.helper"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.with_decorator"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.<top-level>"));
+        // The `@no_type_check` decorator on `free` is folded into its
+        // unit's line range (decorated_definition unwrap).
+        let free_src = doc.blocks.iter().find_map(|b| match b {
+            Block::Code(c) if matches!(&c.common.source_span,
+                SourceSpan::Code{symbol,..} if symbol.as_deref()==Some("kebab_eval.metrics.free")) => Some(c.code.clone()),
+            _ => None,
+        }).unwrap();
+        assert!(free_src.contains("@no_type_check"), "decorator folded in: {free_src}");
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 { assert_eq!(extract_fixture().blocks, a.blocks); }
+    }
+}

crates/kebab-parse-code/src/rust.rs

@@ -30,6 +30,8 @@ use kebab_core::{
 use serde_json::Map;
 use time::OffsetDateTime;
 
+use crate::scaffold::{filename_from_workspace_path, strip_extension};
+
 pub const PARSER_VERSION: &str = "code-rust-v1";
 
 /// Rust AST extractor. Per-unit blocks via tree-sitter-rust 0.24
@@ -162,18 +164,6 @@ impl Extractor for RustAstExtractor {
     }
 }
 
-fn filename_from_workspace_path(p: &str) -> String {
-    p.rsplit('/').next().unwrap_or(p).to_string()
-}
-
-fn strip_extension(filename: &str) -> String {
-    match filename.rfind('.') {
-        Some(0) => filename.to_string(),
-        Some(idx) => filename[..idx].to_string(),
-        None => filename.to_string(),
-    }
-}
-
 fn build_blocks(
     source: &str,
     doc_id: &kebab_core::DocumentId,
@@ -393,7 +383,7 @@ mod tests {
             concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/sample.rs"),
         )
         .unwrap();
-        let asset = kebab_parse_code_test_support::fixed_rust_asset("crates/x/src/sample.rs");
+        let asset = tests_support::fixed_code_asset("crates/x/src/sample.rs", "rust");
         let cfg = kebab_core::ExtractConfig::default();
         let root = std::path::PathBuf::from("/tmp");
         let ctx = kebab_core::ExtractContext { asset: &asset, workspace_root: &root, config: &cfg };
@@ -444,7 +434,7 @@ mod tests {
     /// Run the extractor on an in-memory Rust source string (no fixture
     /// file) and return (symbol, code) for every emitted block.
     fn extract_inline(source: &str) -> Vec<(String, String)> {
-        let asset = kebab_parse_code_test_support::fixed_rust_asset("crates/x/src/inline.rs");
+        let asset = tests_support::fixed_code_asset("crates/x/src/inline.rs", "rust");
         let cfg = kebab_core::ExtractConfig::default();
         let root = std::path::PathBuf::from("/tmp");
         let ctx = kebab_core::ExtractContext { asset: &asset, workspace_root: &root, config: &cfg };
@@ -531,20 +521,23 @@ mod tests {
 }
 
 #[cfg(test)]
-mod kebab_parse_code_test_support {
+pub(crate) mod tests_support {
     use kebab_core::*;
     use time::OffsetDateTime;
-    pub fn fixed_rust_asset(path: &str) -> RawAsset {
+    /// Test-only `RawAsset` builder for any tree-sitter language. Shared
+    /// across `rust.rs` / `python.rs` / future TS+JS extractor tests so all
+    /// in-crate code-extractor tests use a single canonical fixture shape.
+    pub fn fixed_code_asset(workspace_path: &str, code_lang: &str) -> RawAsset {
         RawAsset {
             asset_id: AssetId("a".repeat(64)),
-            source_uri: SourceUri::File(std::path::PathBuf::from(path)),
-            workspace_path: WorkspacePath(path.to_string()),
-            media_type: MediaType::Code("rust".to_string()),
+            source_uri: SourceUri::File(std::path::PathBuf::from(workspace_path)),
+            workspace_path: WorkspacePath(workspace_path.to_string()),
+            media_type: MediaType::Code(code_lang.to_string()),
             byte_len: 0,
             checksum: Checksum("b".repeat(64)),
             discovered_at: OffsetDateTime::from_unix_timestamp(1_700_000_000).unwrap(),
             stored: AssetStorage::Reference {
-                path: std::path::PathBuf::from(path),
+                path: std::path::PathBuf::from(workspace_path),
                 sha: Checksum("b".repeat(64)),
             },
         }

crates/kebab-parse-code/src/scaffold.rs

@@ -0,0 +1,45 @@
+//! `kebab-parse-code::scaffold` — shared pure helpers used by all
+//! per-language extractor modules.
+//!
+//! These are `pub(crate)` utilities extracted from the four extractor
+//! modules (rust / python / typescript / javascript) where identical
+//! copies existed. Keeping them here is the single source of truth.
+
+/// Extract the last path component (filename) from a `/`-separated
+/// workspace path string.
+/// For a path like `crates/x/src/foo.rs` this returns `foo.rs`.
+pub(crate) fn filename_from_workspace_path(p: &str) -> String {
+    p.rsplit('/').next().unwrap_or(p).to_string()
+}
+
+/// Strip the last dot-extension from a filename string.
+/// A leading dot (hidden-file convention) is preserved as-is.
+/// `foo.rs` → `foo`, `.hidden` → `.hidden`, `noext` → `noext`.
+pub(crate) fn strip_extension(filename: &str) -> String {
+    match filename.rfind('.') {
+        Some(0) => filename.to_string(),
+        Some(idx) => filename[..idx].to_string(),
+        None => filename.to_string(),
+    }
+}
+
+/// Join `(mod_prefix, mod_path, name)` into a dotted symbol string.
+///
+/// Used by Python / TypeScript / JavaScript extractors. Rust uses
+/// `::` separators instead and builds symbols inline; this helper
+/// covers the `.`-joined languages.
+///
+/// Empty `mod_prefix` (e.g. file is `__init__.py` at workspace root)
+/// drops the leading prefix segment; empty `mod_path` (file top-level)
+/// drops the class-nesting middle segment.
+pub(crate) fn join_symbol(mod_prefix: &str, mod_path: &[String], name: &str) -> String {
+    let mut parts: Vec<&str> = Vec::with_capacity(mod_path.len() + 2);
+    if !mod_prefix.is_empty() {
+        parts.push(mod_prefix);
+    }
+    for p in mod_path {
+        parts.push(p.as_str());
+    }
+    parts.push(name);
+    parts.join(".")
+}

crates/kebab-parse-code/src/typescript.rs

@@ -0,0 +1,691 @@
+//! `kebab-parse-code::typescript` — tree-sitter TypeScript / TSX AST
+//! extractor (P10-1B Task H).
+//!
+//! Implements [`kebab_core::Extractor`] for [`MediaType::Code("typescript")`].
+//! Walks the tree-sitter parse tree (one of two grammars selected by the
+//! workspace path's extension — `.tsx` uses [`tree_sitter_typescript::LANGUAGE_TSX`],
+//! everything else uses [`tree_sitter_typescript::LANGUAGE_TYPESCRIPT`]) and
+//! emits one [`Block::Code`] per top-level AST semantic unit (free fn,
+//! class, each method, interface, type alias, enum, recursively per
+//! nested class), each carrying [`SourceSpan::Code`] with the unit's
+//! dotted symbol path prefixed by [`module_path_for_tsjs`].
+//!
+//! Glue declarations (`import_statement`, bare `export_statement`
+//! re-exports, `lexical_declaration` / `variable_declaration` at the
+//! module level, namespace / module declarations, etc.) collapse into
+//! one grouped `<top-level>` (or `<module>`) unit.
+//!
+//! `export_statement` is unwrapped: an `export function|class|interface
+//! |type|enum` is treated as the inner declaration arm but the unit's
+//! line range comes from the OUTER `export_statement` so the `export `
+//! prefix is folded in. `export default function () {}` / `export
+//! default class {}` (no `name` field) emits `default` as the symbol
+//! name.
+//!
+//! Scope follows 1A-2 / 1B Task E: AST unit extraction + dotted symbol
+//! paths + line ranges. Per design §3.4 / §9.1 / §9 versioning.
+
+use anyhow::Result;
+use kebab_core::{
+    Block, CanonicalDocument, CodeBlock, CommonBlock, Extractor, Lang, MediaType, Metadata,
+    ParserVersion, Provenance, ProvenanceEvent, ProvenanceKind, SourceSpan, SourceType, TrustLevel,
+    id_for_block, id_for_doc,
+};
+use serde_json::Map;
+use time::OffsetDateTime;
+
+use crate::scaffold::{filename_from_workspace_path, join_symbol, strip_extension};
+
+pub const PARSER_VERSION: &str = "code-ts-v1";
+
+/// TypeScript / TSX AST extractor. Per-unit blocks via
+/// tree-sitter-typescript 0.23 (`LANGUAGE_TYPESCRIPT` / `LANGUAGE_TSX`
+/// — two `LanguageFn`s, selected by extension) parsed by tree-sitter
+/// 0.26.
+pub struct TypescriptAstExtractor;
+
+impl TypescriptAstExtractor {
+    pub fn new() -> Self {
+        Self
+    }
+}
+
+impl Default for TypescriptAstExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl Extractor for TypescriptAstExtractor {
+    fn supports(&self, m: &MediaType) -> bool {
+        matches!(m, MediaType::Code(l) if l == "typescript")
+    }
+
+    fn parser_version(&self) -> ParserVersion {
+        ParserVersion(PARSER_VERSION.to_string())
+    }
+
+    fn extract(
+        &self,
+        ctx: &kebab_core::ExtractContext<'_>,
+        bytes: &[u8],
+    ) -> Result<CanonicalDocument> {
+        let asset = ctx.asset;
+        if !self.supports(&asset.media_type) {
+            anyhow::bail!(
+                "kebab-parse-code: unsupported media_type for TypescriptAstExtractor: {:?}",
+                asset.media_type
+            );
+        }
+
+        let parser_version = self.parser_version();
+        let doc_id = id_for_doc(&asset.workspace_path, &asset.asset_id, &parser_version);
+
+        let source = String::from_utf8(bytes.to_vec()).map_err(|e| {
+            anyhow::anyhow!("kebab-parse-code: TypeScript source is not valid UTF-8: {e}")
+        })?;
+
+        let mod_prefix = crate::lang::module_path_for_tsjs(&asset.workspace_path.0);
+        let language = select_grammar(&asset.workspace_path.0);
+        let blocks = build_blocks(&source, &doc_id, &mod_prefix, language)?;
+        let unit_count = blocks.len() as u32;
+
+        let now = OffsetDateTime::now_utc();
+        let mut events: Vec<ProvenanceEvent> = Vec::with_capacity(2);
+        events.push(ProvenanceEvent {
+            at: asset.discovered_at,
+            agent: "kb-source-fs".to_string(),
+            kind: ProvenanceKind::Discovered,
+            note: None,
+        });
+        events.push(ProvenanceEvent {
+            at: now,
+            agent: "kb-parse-code".to_string(),
+            kind: ProvenanceKind::Parsed,
+            note: Some(format!(
+                "parser_version={}; unit_count={}",
+                parser_version.0, unit_count
+            )),
+        });
+
+        let title = {
+            let fname = filename_from_workspace_path(&asset.workspace_path.0);
+            strip_extension(&fname)
+        };
+
+        // Resolve the file's absolute path for repo detection. If the
+        // source URI carries a relative path, anchor it at the workspace
+        // root so the `.git/` walk-up starts from the right place.
+        let abs_path = match &asset.source_uri {
+            kebab_core::SourceUri::File(p) => {
+                if p.is_absolute() {
+                    p.clone()
+                } else {
+                    ctx.workspace_root.join(p)
+                }
+            }
+            kebab_core::SourceUri::Kb(_) => ctx.workspace_root.to_path_buf(),
+        };
+        let (repo, git_branch, git_commit) = match crate::repo::detect_repo(&abs_path) {
+            Some(r) => (Some(r.name), r.branch, r.commit),
+            None => (None, None, None),
+        };
+
+        let metadata = Metadata {
+            aliases: Vec::new(),
+            tags: Vec::new(),
+            created_at: asset.discovered_at,
+            updated_at: asset.discovered_at,
+            source_type: SourceType::Note,
+            trust_level: TrustLevel::Primary,
+            user_id_alias: None,
+            user: Map::new(),
+            repo,
+            git_branch,
+            git_commit,
+            code_lang: Some("typescript".to_string()),
+        };
+
+        tracing::debug!(
+            target: "kebab-parse-code",
+            "extracted TypeScript doc_id={} workspace_path={} units={}",
+            doc_id.0,
+            asset.workspace_path.0,
+            unit_count
+        );
+
+        Ok(CanonicalDocument {
+            doc_id,
+            source_asset_id: asset.asset_id.clone(),
+            workspace_path: asset.workspace_path.clone(),
+            title,
+            lang: Lang("und".to_string()),
+            blocks,
+            metadata,
+            provenance: Provenance { events },
+            parser_version,
+            schema_version: 1,
+            doc_version: 1,
+            last_chunker_version: None,
+            last_embedding_version: None,
+        })
+    }
+}
+
+/// Select the tree-sitter grammar based on the workspace path's
+/// extension. `.tsx` → TSX grammar; everything else (`.ts`, `.mts`,
+/// `.cts`, `.d.ts`, missing extension) → TypeScript grammar (the JSX-
+/// agnostic variants all share one grammar in tree-sitter-typescript 0.23).
+fn select_grammar(workspace_path: &str) -> tree_sitter::Language {
+    if workspace_path.ends_with(".tsx") {
+        tree_sitter_typescript::LANGUAGE_TSX.into()
+    } else {
+        tree_sitter_typescript::LANGUAGE_TYPESCRIPT.into()
+    }
+}
+
+fn build_blocks(
+    source: &str,
+    doc_id: &kebab_core::DocumentId,
+    mod_prefix: &str,
+    language: tree_sitter::Language,
+) -> anyhow::Result<Vec<kebab_core::Block>> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&language)
+        .map_err(|e| anyhow::anyhow!("set tree-sitter-typescript language: {e}"))?;
+    let tree = parser
+        .parse(source.as_bytes(), None)
+        .ok_or_else(|| anyhow::anyhow!("tree-sitter failed to parse TypeScript source"))?;
+    let lines: Vec<&str> = source.split('\n').collect();
+
+    // units: (symbol, line_start, line_end, is_real_semantic_unit).
+    // Glue groups are pushed with a sentinel symbol + is_real=false so a
+    // post-pass can decide `<module>` vs `<top-level>` (same algorithm
+    // as 1A Gap 1 / 1B Python).
+    let mut units: Vec<(String, u32, u32, bool)> = Vec::new();
+    // (is_module_only_kind 0/1, s, e). `is_module_only_kind` flags
+    // `import_statement` and bare re-export `export_statement`s — used by
+    // the glue flush to pick `<module>` vs `<top-level>` provisional
+    // label (1A's `is_mod_decl` analog).
+    let mut glue: Vec<(usize, u32, u32)> = Vec::new();
+
+    /// Walk preceding `comment` and `decorator` siblings to extend the
+    /// unit's line range upward, folding leading doc/line comments and
+    /// decorators into the unit.
+    ///
+    /// In tree-sitter-typescript 0.23, TS class-method decorators (and
+    /// class-level decorators) are **`class_body` siblings** that
+    /// immediately precede the `method_definition` node — they are NOT
+    /// children of `method_definition`. (Contrast with
+    /// tree-sitter-javascript, where the `decorator` IS stored inside
+    /// `method_definition` as a named child via the `decorator` field, so
+    /// `method_definition.start_row` already covers the decorator line
+    /// there — no sibling walk needed in `javascript.rs`.)
+    ///
+    /// Extending backward over `decorator` siblings here matches Python's
+    /// `decorated_definition` arm behavior: the decorator line is folded
+    /// into the emitted unit's line range.
+    fn unit_start(n: &tree_sitter::Node) -> u32 {
+        let mut start = n.start_position().row as u32 + 1;
+        let mut prev = n.prev_sibling();
+        while let Some(p) = prev {
+            if p.kind() == "comment" || p.kind() == "decorator" {
+                start = p.start_position().row as u32 + 1;
+                prev = p.prev_sibling();
+            } else {
+                break;
+            }
+        }
+        start
+    }
+    fn name_text<'a>(n: &tree_sitter::Node, src: &'a str) -> Option<&'a str> {
+        n.child_by_field_name("name")
+            .map(|c| &src[c.start_byte()..c.end_byte()])
+    }
+    /// Walk a class body, emitting one unit per `method_definition`.
+    /// Class names already pushed onto `mod_path` by the caller, so
+    /// method symbols come out as `<mod_prefix>.<Class>.<method>`.
+    fn walk_class_body(
+        body: tree_sitter::Node,
+        src: &str,
+        mod_prefix: &str,
+        mod_path: &[String],
+        units: &mut Vec<(String, u32, u32, bool)>,
+    ) {
+        let mut cur = body.walk();
+        for child in body.named_children(&mut cur) {
+            if child.kind() == "method_definition" {
+                if let Some(name) = name_text(&child, src) {
+                    let s = unit_start(&child);
+                    let e = child.end_position().row as u32 + 1;
+                    let sym = join_symbol(mod_prefix, mod_path, name);
+                    units.push((sym, s, e, true));
+                }
+            }
+        }
+    }
+    fn walk(
+        node: tree_sitter::Node,
+        src: &str,
+        mod_prefix: &str,
+        mod_path: &[String],
+        units: &mut Vec<(String, u32, u32, bool)>,
+        glue: &mut Vec<(usize, u32, u32)>,
+    ) {
+        let mut cur = node.walk();
+        for child in node.named_children(&mut cur) {
+            let s = unit_start(&child);
+            let e = child.end_position().row as u32 + 1;
+            match child.kind() {
+                "function_declaration" => {
+                    if let Some(name) = name_text(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                    }
+                }
+                "class_declaration" => {
+                    if let Some(name) = name_text(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                        if let Some(body) = child.child_by_field_name("body") {
+                            let mut np = mod_path.to_vec();
+                            np.push(name.to_string());
+                            walk_class_body(body, src, mod_prefix, &np, units);
+                        }
+                    }
+                }
+                "interface_declaration"
+                | "type_alias_declaration"
+                | "enum_declaration" => {
+                    if let Some(name) = name_text(&child, src) {
+                        glue.retain(|(_, gs, _)| *gs < s);
+                        flush_glue(glue, units, mod_prefix, mod_path);
+                        let sym = join_symbol(mod_prefix, mod_path, name);
+                        units.push((sym, s, e, true));
+                    }
+                }
+                "export_statement" => {
+                    // Try field "declaration" first (export class /
+                    // function / interface / type / enum). If absent,
+                    // fall back to "value" — `export default function
+                    // () {}` / `export default class {}` expose the
+                    // anonymous function_expression / class under the
+                    // `value` field (TS grammar 0.23).
+                    let outer_s = s; // includes `export ` prefix line
+                    let outer_e = e;
+                    if let Some(inner) = child.child_by_field_name("declaration") {
+                        let inner_kind = inner.kind();
+                        match inner_kind {
+                            "function_declaration"
+                            | "class_declaration"
+                            | "interface_declaration"
+                            | "type_alias_declaration"
+                            | "enum_declaration" => {
+                                let name_opt = name_text(&inner, src).map(|s| s.to_string());
+                                if let Some(name) = name_opt {
+                                    glue.retain(|(_, gs, _)| *gs < outer_s);
+                                    flush_glue(glue, units, mod_prefix, mod_path);
+                                    let sym =
+                                        join_symbol(mod_prefix, mod_path, &name);
+                                    units.push((sym, outer_s, outer_e, true));
+                                    if inner_kind == "class_declaration" {
+                                        if let Some(body) =
+                                            inner.child_by_field_name("body")
+                                        {
+                                            let mut np = mod_path.to_vec();
+                                            np.push(name);
+                                            walk_class_body(
+                                                body, src, mod_prefix, &np, units,
+                                            );
+                                        }
+                                    }
+                                } else {
+                                    // `export default function foo() {}`
+                                    // path is covered by name_opt =
+                                    // Some(_) above; the no-name path
+                                    // here is `export default` with a
+                                    // function_declaration that
+                                    // somehow lacks `name`. Emit
+                                    // `default` defensively.
+                                    glue.retain(|(_, gs, _)| *gs < outer_s);
+                                    flush_glue(glue, units, mod_prefix, mod_path);
+                                    let sym =
+                                        join_symbol(mod_prefix, mod_path, "default");
+                                    units.push((sym, outer_s, outer_e, true));
+                                }
+                            }
+                            // `lexical_declaration` etc. wrapped in
+                            // export: treat as glue (assigned arrow
+                            // fns / consts don't get their own unit).
+                            _ => {
+                                glue.push((0, s, e));
+                            }
+                        }
+                    } else if let Some(value) = child.child_by_field_name("value") {
+                        // `export default <expr>`. We emit a unit only
+                        // for the function / class shapes (named or
+                        // anonymous); other value shapes are glue.
+                        match value.kind() {
+                            "function_expression"
+                            | "function_declaration"
+                            | "class"
+                            | "class_declaration" => {
+                                let name_opt =
+                                    name_text(&value, src).map(|s| s.to_string());
+                                let leaf = name_opt
+                                    .as_deref()
+                                    .unwrap_or("default")
+                                    .to_string();
+                                glue.retain(|(_, gs, _)| *gs < outer_s);
+                                flush_glue(glue, units, mod_prefix, mod_path);
+                                let sym = join_symbol(mod_prefix, mod_path, &leaf);
+                                units.push((sym, outer_s, outer_e, true));
+                                // Recurse into class body if we have one.
+                                if matches!(
+                                    value.kind(),
+                                    "class" | "class_declaration"
+                                ) {
+                                    if let Some(body) =
+                                        value.child_by_field_name("body")
+                                    {
+                                        let mut np = mod_path.to_vec();
+                                        np.push(leaf);
+                                        walk_class_body(
+                                            body, src, mod_prefix, &np, units,
+                                        );
+                                    }
+                                }
+                            }
+                            _ => {
+                                glue.push((0, s, e));
+                            }
+                        }
+                    } else {
+                        // Bare `export { x };` / `export * from "..."` —
+                        // a re-export, glue with module-only flag set
+                        // (we have no `declaration` / `value` field for
+                        // it).
+                        glue.push((1, s, e));
+                    }
+                }
+                "import_statement" => {
+                    glue.push((1, s, e));
+                }
+                "lexical_declaration" | "variable_declaration" => {
+                    glue.push((0, s, e));
+                }
+                // Namespace / module declarations (rare in app code,
+                // common in `.d.ts`): treat as glue per plan §Task H
+                // (1B 1차 scope; documented under spec Risks).
+                "internal_module" | "module" | "ambient_declaration" => {
+                    glue.push((0, s, e));
+                }
+                _ => {}
+            }
+        }
+        flush_glue(glue, units, mod_prefix, mod_path);
+    }
+    fn flush_glue(
+        glue: &mut Vec<(usize, u32, u32)>,
+        units: &mut Vec<(String, u32, u32, bool)>,
+        mod_prefix: &str,
+        mod_path: &[String],
+    ) {
+        if glue.is_empty() {
+            return;
+        }
+        let s = glue.iter().map(|(_, a, _)| *a).min().unwrap();
+        let e = glue.iter().map(|(_, _, b)| *b).max().unwrap();
+        let only_module = glue.iter().all(|(is_mod, _, _)| *is_mod == 1);
+        let label = if only_module { "<module>" } else { "<top-level>" };
+        units.push((join_symbol(mod_prefix, mod_path, label), s, e, false));
+        glue.clear();
+    }
+
+    walk(
+        tree.root_node(),
+        source,
+        mod_prefix,
+        &[],
+        &mut units,
+        &mut glue,
+    );
+
+    // `<module>` is correct only when the file produced no real unit.
+    // Otherwise the import-only group becomes `<top-level>` (same
+    // post-pass as 1A Gap 1 / Python).
+    let has_real_unit = units.iter().any(|(_, _, _, is_real)| *is_real);
+    if has_real_unit {
+        for (sym, _, _, is_real) in units.iter_mut() {
+            if !*is_real && sym.ends_with("<module>") {
+                let pre = &sym[..sym.len() - "<module>".len()];
+                *sym = format!("{pre}<top-level>");
+            }
+        }
+    }
+
+    let total_lines = lines.len() as u32;
+    let mut blocks = Vec::with_capacity(units.len());
+    for (ordinal, (symbol, ls, le, _is_real)) in units.into_iter().enumerate() {
+        let line_start = ls.max(1);
+        let line_end = le.min(total_lines.max(1));
+        let span = SourceSpan::Code {
+            line_start,
+            line_end,
+            symbol: Some(symbol),
+            lang: Some("typescript".to_string()),
+        };
+        let block_id = id_for_block(doc_id, "code", &[], ordinal as u32, &span);
+        let code = lines[(line_start as usize - 1)..=(line_end as usize - 1)].join("\n");
+        blocks.push(Block::Code(CodeBlock {
+            common: CommonBlock {
+                block_id,
+                heading_path: Vec::new(),
+                source_span: span,
+            },
+            lang: Some("typescript".to_string()),
+            code,
+        }));
+    }
+    Ok(blocks)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture(name: &str, workspace_path: &str) -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(format!(
+            concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/{}"),
+            name
+        ))
+        .unwrap();
+        let asset = crate::rust::tests_support::fixed_code_asset(workspace_path, "typescript");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        TypescriptAstExtractor::new()
+            .extract(&ctx, &bytes)
+            .unwrap()
+    }
+
+    fn symbols(doc: &kebab_core::CanonicalDocument) -> Vec<String> {
+        let mut s: Vec<String> = doc
+            .blocks
+            .iter()
+            .filter_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, lang, .. } => {
+                        assert_eq!(lang.as_deref(), Some("typescript"));
+                        symbol.clone()
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .collect();
+        s.sort();
+        s
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_typescript() {
+        let e = TypescriptAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("typescript".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+
+    #[test]
+    fn ts_units_match_design_3_4_symbols() {
+        // workspace_path `src/sample.ts` → mod_prefix `src/sample`
+        let doc = extract_fixture("sample.ts", "src/sample.ts");
+        let syms = symbols(&doc);
+        assert!(syms.iter().any(|s| s == "src/sample.add"), "got {syms:?}");
+        assert!(syms.iter().any(|s| s == "src/sample.Greet"));
+        assert!(syms.iter().any(|s| s == "src/sample.Maybe"));
+        assert!(syms.iter().any(|s| s == "src/sample.Retriever"));
+        assert!(syms.iter().any(|s| s == "src/sample.Retriever.search"));
+        assert!(syms.iter().any(|s| s == "src/sample.Retriever.create"));
+        assert!(syms.iter().any(|s| s == "src/sample.default"));
+        assert!(syms.iter().any(|s| s == "src/sample.<top-level>"));
+    }
+
+    #[test]
+    fn tsx_uses_tsx_grammar_and_emits_units() {
+        let doc = extract_fixture("sample.tsx", "src/sample.tsx");
+        let syms = symbols(&doc);
+        assert!(
+            syms.iter().any(|s| s == "src/sample.Hello"),
+            "got {syms:?}"
+        );
+        assert!(
+            syms.iter().any(|s| s == "src/sample.<top-level>"),
+            "arrow fn + import should roll into top-level glue"
+        );
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture("sample.ts", "src/sample.ts");
+        for _ in 0..30 {
+            assert_eq!(extract_fixture("sample.ts", "src/sample.ts").blocks, a.blocks);
+        }
+    }
+
+    /// Regression: TS class-method decorators are `class_body` preceding
+    /// siblings (not children of `method_definition`). The `unit_start`
+    /// backward walk must fold the decorator line into the emitted unit's
+    /// line range, matching Python's `decorated_definition` behavior.
+    #[test]
+    fn class_method_decorator_folded_into_method_unit() {
+        // Line 1 (1-indexed): "class Foo {"
+        // Line 2:             "    @Log()"   <- decorator
+        // Line 3:             "    bar() { return 1; }"
+        // Line 4:             "}"
+        let bytes = b"class Foo {\n    @Log()\n    bar() { return 1; }\n}\n";
+        let asset = crate::rust::tests_support::fixed_code_asset("src/foo.ts", "typescript");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        let doc = TypescriptAstExtractor::new().extract(&ctx, bytes).unwrap();
+
+        let bar_block = doc
+            .blocks
+            .iter()
+            .find_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, .. }
+                        if symbol.as_deref() == Some("src/foo.Foo.bar") =>
+                    {
+                        Some(c)
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .expect("src/foo.Foo.bar block should be present");
+
+        // After the fix, the unit MUST include the @Log() decorator line.
+        assert!(
+            bar_block.code.contains("@Log()"),
+            "decorator must be folded into class-method unit (Python parity); got code: {:?}",
+            bar_block.code
+        );
+
+        // line_start must be 2 (the @Log() line), NOT 3 (the bar() line).
+        match &bar_block.common.source_span {
+            SourceSpan::Code { line_start, .. } => {
+                assert_eq!(
+                    *line_start, 2,
+                    "line_start must cover the @Log() decorator line (got {line_start})"
+                );
+            }
+            _ => unreachable!(),
+        }
+    }
+
+    /// Class-level decorator (preceding sibling of `class_declaration` in
+    /// the module root): same `unit_start` backward walk folds it in.
+    /// Line 1: "@Injectable()"
+    /// Line 2: "class Service {"
+    /// Line 3: "}"
+    #[test]
+    fn ts_class_decorator_folded_into_class_unit() {
+        let bytes = b"@Injectable()\nclass Service {\n}\n";
+        let asset = crate::rust::tests_support::fixed_code_asset("src/svc.ts", "typescript");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext {
+            asset: &asset,
+            workspace_root: &root,
+            config: &cfg,
+        };
+        let doc = TypescriptAstExtractor::new().extract(&ctx, bytes).unwrap();
+
+        let svc_block = doc
+            .blocks
+            .iter()
+            .find_map(|b| match b {
+                Block::Code(c) => match &c.common.source_span {
+                    SourceSpan::Code { symbol, .. }
+                        if symbol.as_deref() == Some("src/svc.Service") =>
+                    {
+                        Some(c)
+                    }
+                    _ => None,
+                },
+                _ => None,
+            })
+            .expect("src/svc.Service block should be present");
+
+        assert!(
+            svc_block.code.contains("@Injectable()"),
+            "class-level decorator must be folded into the class unit; got code: {:?}",
+            svc_block.code
+        );
+        match &svc_block.common.source_span {
+            SourceSpan::Code { line_start, .. } => {
+                assert_eq!(
+                    *line_start, 1,
+                    "line_start must cover the @Injectable() line (got {line_start})"
+                );
+            }
+            _ => unreachable!(),
+        }
+    }
+}

crates/kebab-parse-code/tests/fixtures/sample.go

@@ -0,0 +1,34 @@
+// sample.go
+package chunk
+
+import (
+	"fmt"
+	"strings"
+)
+
+const Version = "v1"
+
+type MdHeadingV1Chunker struct {
+	Name string
+}
+
+// ChunkDoc returns a stub list of strings.
+func (m *MdHeadingV1Chunker) ChunkDoc(input string) []string {
+	return []string{m.Name}
+}
+
+func (m MdHeadingV1Chunker) Name2() string {
+	return m.Name
+}
+
+type Stringer interface {
+	String() string
+}
+
+func Free(x int) int {
+	return x + 1
+}
+
+func init() {
+	fmt.Println(strings.ToUpper("init"))
+}

crates/kebab-parse-code/tests/fixtures/sample.java

@@ -0,0 +1,36 @@
+// sample.java
+package com.kebab.chunk;
+
+import java.util.List;
+import java.util.stream.Collectors;
+
+/**
+ * Heading-aware Markdown chunker.
+ */
+public class MdHeadingV1Chunker {
+    private final String name;
+
+    public MdHeadingV1Chunker(String name) {
+        this.name = name;
+    }
+
+    public List<String> chunkDoc(String input) {
+        return List.of(name, input);
+    }
+
+    public String getName() {
+        return name;
+    }
+
+    public static class Builder {
+        private String name;
+        public Builder withName(String n) { this.name = n; return this; }
+        public MdHeadingV1Chunker build() { return new MdHeadingV1Chunker(name); }
+    }
+}
+
+interface Stringer {
+    String asString();
+}
+
+enum Mode { DEFAULT, FAST }

crates/kebab-parse-code/tests/fixtures/sample.js

@@ -0,0 +1,9 @@
+// sample.js
+import { x } from "./other";
+const ANSWER = 42;
+export function add(a, b) { return a + b; }
+export class Retriever {
+    search(q) { return []; }
+    static create() { return new Retriever(); }
+}
+export default function () { return 1; }

crates/kebab-parse-code/tests/fixtures/sample.kt

@@ -0,0 +1,29 @@
+// sample.kt
+package com.kebab.chunk
+
+import java.util.List
+
+/**
+ * Heading-aware Markdown chunker.
+ */
+class MdHeadingV1Chunker(val name: String) {
+    fun chunkDoc(input: String): List<String> = listOf(name, input)
+
+    fun getName(): String = name
+
+    companion object {
+        fun withName(n: String): MdHeadingV1Chunker = MdHeadingV1Chunker(n)
+    }
+}
+
+interface Stringer {
+    fun asString(): String
+}
+
+enum class Mode { DEFAULT, FAST }
+
+fun freeFunction(x: Int): Int = x + 1
+
+object Singleton {
+    fun ping(): String = "pong"
+}

crates/kebab-parse-code/tests/fixtures/sample.py

@@ -0,0 +1,26 @@
+"""sample fixture."""
+import os
+
+ANSWER = 42
+
+@no_type_check
+def free(x):
+    """free fn."""
+    return x + 1
+
+class Foo:
+    """doc."""
+    def double(self, n):
+        return n * 2
+
+    @classmethod
+    def name(cls):
+        return "foo"
+
+class Outer:
+    class Inner:
+        def helper(self):
+            return True
+
+def with_decorator():
+    pass

crates/kebab-parse-code/tests/fixtures/sample.ts

@@ -0,0 +1,11 @@
+// sample.ts
+import { x } from "./other";
+const ANSWER = 42;
+export interface Greet { hello(): string; }
+export type Maybe<T> = T | null;
+export function add(a: number, b: number): number { return a + b; }
+export class Retriever {
+    search(q: string): string[] { return []; }
+    static create(): Retriever { return new Retriever(); }
+}
+export default function () { return 1; }

crates/kebab-parse-code/tests/fixtures/sample.tsx

@@ -0,0 +1,4 @@
+// sample.tsx
+import React from "react";
+export function Hello({ name }: { name: string }) { return <span>{name}</span>; }
+export const App = () => <Hello name="x" />;  // arrow fn assigned → glue

crates/kebab-parse-code/tests/lang.rs

@@ -9,6 +9,8 @@ fn known_extensions_map_to_canonical_identifiers() {
         ("foo.pyi", Some("python")),
         ("foo.ts", Some("typescript")),
         ("foo.tsx", Some("typescript")),
+        ("foo.mts", Some("typescript")),  // ESM TS — same grammar
+        ("foo.cts", Some("typescript")),  // CommonJS TS — same grammar
         ("foo.js", Some("javascript")),
         ("foo.mjs", Some("javascript")),
         ("foo.cjs", Some("javascript")),

crates/kebab-search/src/lexical.rs

@@ -346,6 +346,34 @@ fn run_query(
         }
     }
 
+    // p10-1A-1 fix (dogfood-discovered 2026-05-20): code_lang filter
+    // (IN-list on metadata_json.$.code_lang). Empty Vec = no filter.
+    if !filters.code_lang.is_empty() {
+        let placeholders = std::iter::repeat_n("?", filters.code_lang.len())
+            .collect::<Vec<_>>()
+            .join(",");
+        sql.push_str(&format!(
+            " AND json_extract(d.metadata_json, '$.code_lang') IN ({placeholders})"
+        ));
+        for lang in &filters.code_lang {
+            params.push(Box::new(lang.clone()));
+        }
+    }
+
+    // p10-1A-1 fix (dogfood-discovered 2026-05-20): repo filter
+    // (IN-list on metadata_json.$.repo). Empty Vec = no filter.
+    if !filters.repo.is_empty() {
+        let placeholders = std::iter::repeat_n("?", filters.repo.len())
+            .collect::<Vec<_>>()
+            .join(",");
+        sql.push_str(&format!(
+            " AND json_extract(d.metadata_json, '$.repo') IN ({placeholders})"
+        ));
+        for repo in &filters.repo {
+            params.push(Box::new(repo.clone()));
+        }
+    }
+
     // p9-fb-36: ingested_after filter.
     // `documents.updated_at` is RFC3339 stored as TEXT (always UTC `Z` per
     // fb-32 ingest path), so lexicographic >= compare is correct — but only

crates/kebab-search/tests/lexical.rs

@@ -785,6 +785,19 @@ impl TestEnv {
         body: &str,
         media: MediaType,
         updated_at: OffsetDateTime,
+    ) -> DocumentId {
+        self.insert_doc_full_with_metadata(path, body, media, updated_at, "{}")
+    }
+
+    /// Like `insert_doc_full` but accepts an explicit `metadata_json` string
+    /// so p10-1A-1 filter tests can set `metadata.code_lang` / `metadata.repo`.
+    fn insert_doc_full_with_metadata(
+        &self,
+        path: &str,
+        body: &str,
+        media: MediaType,
+        updated_at: OffsetDateTime,
+        metadata_json: &str,
     ) -> DocumentId {
         use time::format_description::well_known::Rfc3339;
         let doc_id = self.next_id("doc");
@@ -810,10 +823,10 @@ impl TestEnv {
                 source_type, trust_level, parser_version,
                 doc_version, schema_version, metadata_json,
                 provenance_json, created_at, updated_at
-            ) VALUES (?, ?, ?, NULL, 'en', 'markdown', 'primary', 'pv1', 1, 1,
-                      '{}', '{\"events\":[]}',
+            ) VALUES (?, ?, ?, NULL, 'en', 'code', 'primary', 'pv1', 1, 1,
+                      ?, '{\"events\":[]}',
                       '2024-01-01T00:00:00Z', ?)",
-            rusqlite::params![doc_id, asset_id, path, updated_at_str],
+            rusqlite::params![doc_id, asset_id, path, metadata_json, updated_at_str],
         )
         .expect("insert document");
 
@@ -834,6 +847,21 @@ impl TestEnv {
         DocumentId(doc_id)
     }
 
+    /// Insert a code doc with explicit `code_lang` and optional `repo` in metadata.
+    fn insert_code_doc(&self, path: &str, body: &str, code_lang: &str, repo: Option<&str>) -> DocumentId {
+        let metadata_json = match repo {
+            Some(r) => format!(r#"{{"code_lang":"{code_lang}","repo":"{r}"}}"#),
+            None => format!(r#"{{"code_lang":"{code_lang}"}}"#),
+        };
+        self.insert_doc_full_with_metadata(
+            path,
+            body,
+            MediaType::Markdown,
+            OffsetDateTime::now_utc(),
+            &metadata_json,
+        )
+    }
+
     fn run_search(&self, query: &str, filters: &SearchFilters) -> Vec<SearchHit> {
         let r = self.inner.retriever();
         let q = SearchQuery {
@@ -934,6 +962,52 @@ fn lexical_empty_filters_match_default_behavior() {
     assert!(!with_default.is_empty());
 }
 
+// ── p10-1A-1 filter tests ────────────────────────────────────────────────
+
+#[test]
+fn lexical_filter_by_code_lang() {
+    // Three docs: python code, rust code, markdown (no code_lang).
+    // Filter code_lang=["python"] → only the python doc should match.
+    let env = TestEnv::new();
+    env.insert_code_doc("src/main.py", "AsyncClient session", "python", None);
+    env.insert_code_doc("src/lib.rs", "AsyncClient session", "rust", None);
+    env.insert_doc("docs/guide.md", "AsyncClient session");
+
+    let filters = SearchFilters {
+        code_lang: vec!["python".to_string()],
+        ..Default::default()
+    };
+    let hits = env.run_search("AsyncClient", &filters);
+    assert_eq!(hits.len(), 1, "only python doc should match code_lang filter");
+    assert!(
+        hits[0].doc_path.0.ends_with(".py"),
+        "expected python path, got: {}",
+        hits[0].doc_path.0
+    );
+}
+
+#[test]
+fn lexical_filter_by_repo() {
+    // Three docs: one in repo "httpx", one in repo "requests", one with no repo.
+    // Filter repo=["httpx"] → only the httpx doc should match.
+    let env = TestEnv::new();
+    env.insert_code_doc("httpx/client.py", "session send request", "python", Some("httpx"));
+    env.insert_code_doc("requests/api.py", "session send request", "python", Some("requests"));
+    env.insert_code_doc("standalone.py", "session send request", "python", None);
+
+    let filters = SearchFilters {
+        repo: vec!["httpx".to_string()],
+        ..Default::default()
+    };
+    let hits = env.run_search("session", &filters);
+    assert_eq!(hits.len(), 1, "only httpx doc should match repo filter");
+    assert!(
+        hits[0].doc_path.0.starts_with("httpx/"),
+        "expected httpx path, got: {}",
+        hits[0].doc_path.0
+    );
+}
+
 #[test]
 fn lexical_snapshot_run_1() {
     // Pinned snapshot. A small, deterministic corpus; the JSON shape of

crates/kebab-source-fs/Cargo.toml

@@ -18,6 +18,7 @@ blake3       = { workspace = true }
 tracing      = { workspace = true }
 walkdir      = "2"
 ignore       = "0.4"
+globset      = "0.4"
 
 [dev-dependencies]
 serde_json   = { workspace = true }

crates/kebab-source-fs/src/connector.rs

@@ -86,7 +86,7 @@ impl FsSourceConnector {
         excludes.extend(scope.exclude.iter().cloned());
         let kbignore = read_kbignore(&root)?;
 
-        let overrides = build_overrides(&root, &excludes, &kbignore)?;
+        let overrides = build_overrides(&root, &excludes, &kbignore, &scope.include)?;
         Ok((root, overrides))
     }
 
@@ -103,8 +103,6 @@ impl FsSourceConnector {
     ) -> Result<(Vec<RawAsset>, FsScanSkips)> {
         let (root, overrides) = self.resolve_scan_params(scope)?;
 
-        log_scope_include_warning(scope);
-
         let (files, skipped_entries) = walk_files_with_skips(&root, &overrides)?;
 
         // Accumulate per-category skip counts and sample paths.
@@ -284,14 +282,6 @@ fn build_assets(
     Ok(assets)
 }
 
-fn log_scope_include_warning(scope: &SourceScope) {
-    if !scope.include.is_empty() {
-        tracing::debug!(
-            count = scope.include.len(),
-            "FsSourceConnector ignores scope.include — handled by extractor router"
-        );
-    }
-}
 
 impl SourceConnector for FsSourceConnector {
     fn scan(&self, scope: &SourceScope) -> Result<Vec<RawAsset>> {

crates/kebab-source-fs/src/media.rs

@@ -19,7 +19,9 @@ pub(crate) fn media_type_for(path: &Path) -> MediaType {
         .unwrap_or_default();
 
     match ext.as_str() {
-        "md" => MediaType::Markdown,
+        // Markdown + MDX (markdown + JSX, treated as plain markdown — the
+        // JSX islands are folded into raw passthrough by the md parser).
+        "md" | "mdx" => MediaType::Markdown,
         "pdf" => MediaType::Pdf,
 
         "png" => MediaType::Image(ImageType::Png),
@@ -38,6 +40,19 @@ pub(crate) fn media_type_for(path: &Path) -> MediaType {
         // recognized code langs stay Other until their phase (1B+).
         "rs" => MediaType::Code("rust".to_string()),
 
+        // p10-1B: Python / TS / JS AST chunkers active.
+        "py" | "pyi"               => MediaType::Code("python".into()),
+        // .mts / .cts are TypeScript ESM / CommonJS variants — same grammar.
+        "ts" | "tsx" | "mts" | "cts" => MediaType::Code("typescript".into()),
+        "js" | "mjs" | "cjs" | "jsx" => MediaType::Code("javascript".into()),
+
+        // p10-1C-Go: Go ingest activated.
+        "go" => MediaType::Code("go".into()),
+
+        // p10-1C-JK: JVM family (Java + Kotlin) ingest activated.
+        "java"             => MediaType::Code("java".into()),
+        "kt" | "kts"       => MediaType::Code("kotlin".into()),
+
         // Empty string (no extension) and any other extension: bucket as
         // Other and let downstream extractors decide if they support it.
         _ => MediaType::Other(ext),
@@ -81,11 +96,48 @@ mod tests {
             media_type_for(Path::new("crates/kebab-core/src/lib.rs")),
             MediaType::Code("rust".to_string())
         );
-        // non-Rust code extensions stay Other in 1A
-        assert_eq!(media_type_for(Path::new("a/b.py")), MediaType::Other("py".to_string()));
         assert_eq!(media_type_for(Path::new("Cargo.toml")), MediaType::Other("toml".to_string()));
     }
 
+    #[test]
+    fn py_ts_js_files_map_to_media_code() {
+        assert_eq!(media_type_for(Path::new("a/b.py")),    MediaType::Code("python".into()));
+        assert_eq!(media_type_for(Path::new("a/b.pyi")),   MediaType::Code("python".into()));
+        assert_eq!(media_type_for(Path::new("a/b.ts")),    MediaType::Code("typescript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.tsx")),   MediaType::Code("typescript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.js")),    MediaType::Code("javascript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.mjs")),   MediaType::Code("javascript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.cjs")),   MediaType::Code("javascript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.jsx")),   MediaType::Code("javascript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.rs")),    MediaType::Code("rust".into()));
+    }
+
+    #[test]
+    fn ts_variants_mts_cts() {
+        // .mts / .cts are TypeScript ESM / CommonJS — same grammar as .ts.
+        assert_eq!(media_type_for(Path::new("a/b.mts")), MediaType::Code("typescript".into()));
+        assert_eq!(media_type_for(Path::new("a/b.cts")), MediaType::Code("typescript".into()));
+    }
+
+    #[test]
+    fn mdx_routes_to_markdown() {
+        // MDX is markdown with JSX islands; the md parser folds the JSX
+        // through as raw passthrough.
+        assert_eq!(media_type_for(Path::new("docs/page.mdx")), MediaType::Markdown);
+    }
+
+    #[test]
+    fn go_files_map_to_media_code_go() {
+        assert_eq!(media_type_for(Path::new("a/b.go")), MediaType::Code("go".into()));
+    }
+
+    #[test]
+    fn java_kotlin_files_map_to_media_code() {
+        assert_eq!(media_type_for(Path::new("a/b.java")), MediaType::Code("java".into()));
+        assert_eq!(media_type_for(Path::new("a/b.kt")), MediaType::Code("kotlin".into()));
+        assert_eq!(media_type_for(Path::new("a/b.kts")), MediaType::Code("kotlin".into()));
+    }
+
     #[test]
     fn unknown_and_missing_extension() {
         assert_eq!(

crates/kebab-source-fs/src/walker.rs

@@ -44,6 +44,7 @@ use std::collections::HashSet;
 use std::path::{Path, PathBuf};
 
 use anyhow::{Context, Result};
+use globset::{GlobBuilder, GlobSet, GlobSetBuilder};
 use ignore::overrides::{Override, OverrideBuilder};
 use walkdir::WalkDir;
 
@@ -69,6 +70,11 @@ const DEFAULT_EXCLUDES: &[&str] = &[
 ///
 /// `default_and_config` covers DEFAULT_EXCLUDES + `config.workspace.exclude`
 /// — these do NOT map to any of the three named `IngestReport` counters.
+///
+/// `include` is the compiled `scope.include` allow-list. When the set is
+/// empty (no patterns) every file passes; when non-empty a file must match
+/// at least one pattern to be accepted (directories always pass, so the
+/// walker can still descend into them).
 pub(crate) struct WalkOverrides {
     /// Merged matcher — same as today's `Override`; used for the walk decision.
     pub combined: Override,
@@ -78,6 +84,8 @@ pub(crate) struct WalkOverrides {
     pub kebabignore: Override,
     /// Matcher built from `kebab_parse_code::BUILTIN_BLACKLIST` only.
     pub builtin: Override,
+    /// Compiled allow-list from `scope.include`. Empty set = pass all.
+    pub include: GlobSet,
 }
 
 /// Skip attribution category. Used by the connector when counting per-source
@@ -161,10 +169,15 @@ fn build_single_matcher_owned(root: &Path, patterns: &[String]) -> Result<Overri
 /// The three per-source matchers (`gitignore`, `kebabignore`, `builtin`) are
 /// built in addition to the combined one so the connector can attribute skips
 /// to the correct `IngestReport` counter without a second walker pass.
+///
+/// `include_patterns` (from `scope.include`) are compiled into an allow-list
+/// `GlobSet`. Empty slice → pass-all (backward-compat); non-empty → file
+/// must match at least one pattern to be accepted.
 pub(crate) fn build_overrides(
     root: &Path,
     config_exclude: &[String],
     kbignore_patterns: &[String],
+    include_patterns: &[String],
 ) -> Result<WalkOverrides> {
     let gitignore_patterns = read_gitignore(root)?;
 
@@ -209,14 +222,41 @@ pub(crate) fn build_overrides(
         .build()
         .context("failed to compile combined override set")?;
 
+    // Allow-list GlobSet: empty Vec → matches nothing (= pass all); non-empty
+    // → file must match at least one glob to be accepted. We compile with
+    // `case_insensitive=false` to keep the semantics consistent with the
+    // OverrideBuilder exclude patterns above.
+    let include = build_include_globset(include_patterns)?;
+
     Ok(WalkOverrides {
         combined,
         gitignore,
         kebabignore,
         builtin,
+        include,
     })
 }
 
+/// Compile `scope.include` patterns into a `GlobSet` allow-list.
+///
+/// Each pattern uses `GlobBuilder` with `literal_separator = true` so that
+/// `**` can cross directory boundaries while `*` stops at `/`, matching the
+/// gitignore convention used throughout the rest of the walker.
+///
+/// An empty slice produces an empty `GlobSet` — callers interpret that as
+/// "pass all files" (no allow-list constraint).
+fn build_include_globset(patterns: &[String]) -> Result<GlobSet> {
+    let mut builder = GlobSetBuilder::new();
+    for pat in patterns {
+        let glob = GlobBuilder::new(pat)
+            .literal_separator(true)
+            .build()
+            .with_context(|| format!("invalid include pattern: {pat}"))?;
+        builder.add(glob);
+    }
+    builder.build().context("failed to compile include globset")
+}
+
 /// Classify why a path was excluded, using per-source matchers in spec §5.2
 /// priority order: built-in > gitignore > kebabignore > other.
 ///
@@ -391,6 +431,13 @@ pub(crate) fn walk_files_with_skips(
         }
 
         if entry.file_type().is_file() {
+            // Apply include allow-list: if non-empty, the file's path
+            // relative to root must match at least one pattern.
+            if !overrides.include.is_empty() && !overrides.include.is_match(rel) {
+                // Not in the allow-list — silently drop (no skip counter;
+                // the include filter is not a "skip" source in IngestReport).
+                continue;
+            }
             accepted.push(path.to_path_buf());
         }
     }
@@ -406,7 +453,7 @@ mod tests {
     #[test]
     fn empty_inputs_compile_into_an_override() {
         let dir = tempfile::tempdir().unwrap();
-        let ov = build_overrides(dir.path(), &[], &[]).unwrap();
+        let ov = build_overrides(dir.path(), &[], &[], &[]).unwrap();
         // Default-excludes only; non-special files should not match.
         let m = ov.combined.matched(Path::new("notes/alpha.md"), false);
         assert!(!m.is_ignore());
@@ -415,7 +462,7 @@ mod tests {
     #[test]
     fn default_excludes_ds_store_and_resource_forks() {
         let dir = tempfile::tempdir().unwrap();
-        let ov = build_overrides(dir.path(), &[], &[]).unwrap();
+        let ov = build_overrides(dir.path(), &[], &[], &[]).unwrap();
         assert!(ov.combined.matched(Path::new(".DS_Store"), false).is_ignore());
         assert!(
             ov.combined.matched(Path::new("notes/.DS_Store"), false).is_ignore()
@@ -433,6 +480,7 @@ mod tests {
             dir.path(),
             &["*.tmp".to_string(), "node_modules/**".to_string()],
             &[],
+            &[],
         )
         .unwrap();
         assert!(ov.combined.matched(Path::new("a.tmp"), false).is_ignore());
@@ -452,6 +500,7 @@ mod tests {
             dir.path(),
             &["*.tmp".to_string()],
             &["secret/**".to_string()],
+            &[],
         )
         .unwrap();
         assert!(ov.combined.matched(Path::new("a.tmp"), false).is_ignore());
@@ -491,7 +540,7 @@ mod tests {
         fs::write(root.join("src/main.rs"), "x").unwrap();
         fs::write(root.join("node_modules/foo/bar.js"), "x").unwrap();
 
-        let overrides = build_overrides(root, &[], &[]).unwrap();
+        let overrides = build_overrides(root, &[], &[], &[]).unwrap();
         // Override::matched expects paths relative to the builder's root.
         let m_in = overrides.combined.matched(Path::new("src/main.rs"), false);
         let m_out = overrides.combined.matched(Path::new("node_modules/foo/bar.js"), false);
@@ -514,7 +563,7 @@ mod tests {
         fs::create_dir_all(root.join("ok")).unwrap();
         fs::write(root.join("ok/z.txt"), "z").unwrap();
 
-        let overrides = build_overrides(root, &[], &[]).unwrap();
+        let overrides = build_overrides(root, &[], &[], &[]).unwrap();
         // Override::matched expects paths relative to the builder's root.
         for blacklisted in [
             "target/x/y.txt",
@@ -544,7 +593,7 @@ mod tests {
         fs::create_dir_all(root.join("dist")).unwrap();
         fs::write(root.join("dist/bundle.js"), "x").unwrap();
 
-        let overrides = build_overrides(root, &[], &[]).unwrap();
+        let overrides = build_overrides(root, &[], &[], &[]).unwrap();
         assert!(overrides.combined.matched(Path::new("a.log"), false).is_ignore());
         assert!(overrides.combined.matched(Path::new("dist/bundle.js"), false).is_ignore());
         assert!(!overrides.combined.matched(Path::new("src/main.rs"), false).is_ignore());
@@ -562,7 +611,7 @@ mod tests {
         fs::write(root.join("src/main.rs"), "x").unwrap();
 
         // No .gitignore present — patterns from .gitignore should not affect overrides.
-        let overrides = build_overrides(root, &[], &[]).unwrap();
+        let overrides = build_overrides(root, &[], &[], &[]).unwrap();
         assert!(!overrides.combined.matched(Path::new("a.log"), false).is_ignore());
         assert!(!overrides.combined.matched(Path::new("src/main.rs"), false).is_ignore());
     }
@@ -577,7 +626,7 @@ mod tests {
         // semantics, but at minimum it must not produce double-`!` corruption.
         fs::write(root.join(".gitignore"), "!keep/\n").unwrap();
         // Just verify build_overrides doesn't error.
-        let result = build_overrides(root, &[], &[]);
+        let result = build_overrides(root, &[], &[], &[]);
         assert!(result.is_ok(), "should not error on negation pattern: {:?}", result.err());
     }
 
@@ -594,7 +643,7 @@ mod tests {
         // .gitignore entry. Builtin must win (priority order §5.2).
         fs::write(root.join(".gitignore"), "node_modules/\n").unwrap();
 
-        let ov = build_overrides(root, &[], &[]).unwrap();
+        let ov = build_overrides(root, &[], &[], &[]).unwrap();
         // node_modules/ dir itself
         let cat = classify_skip(Path::new("node_modules"), true, &ov);
         assert_eq!(cat, SkipCategory::BuiltinBlacklist, "builtin must have priority");
@@ -609,7 +658,7 @@ mod tests {
         let root = tmp.path();
         fs::write(root.join(".gitignore"), "*.log\n").unwrap();
 
-        let ov = build_overrides(root, &[], &[]).unwrap();
+        let ov = build_overrides(root, &[], &[], &[]).unwrap();
         let cat = classify_skip(Path::new("app.log"), false, &ov);
         assert_eq!(cat, SkipCategory::Gitignore);
     }
@@ -621,7 +670,7 @@ mod tests {
         let tmp = TempDir::new().unwrap();
         let root = tmp.path();
 
-        let ov = build_overrides(root, &[], &["*.secret".to_string()]).unwrap();
+        let ov = build_overrides(root, &[], &["*.secret".to_string()], &[]).unwrap();
         let cat = classify_skip(Path::new("creds.secret"), false, &ov);
         assert_eq!(cat, SkipCategory::Kebabignore);
     }
@@ -637,7 +686,7 @@ mod tests {
         fs::write(root.join("ok.md"), "# ok").unwrap();
         fs::write(root.join("skipme.log"), "x").unwrap();
 
-        let ov = build_overrides(root, &[], &[]).unwrap();
+        let ov = build_overrides(root, &[], &[], &[]).unwrap();
         let (accepted, skipped_entries) = walk_files_with_skips(root, &ov).unwrap();
 
         let accepted_names: Vec<_> = accepted
@@ -677,7 +726,7 @@ mod tests {
         fs::write(root.join("node_modules/foo/bar.js"), "x").unwrap();
         fs::write(root.join("ok.md"), "# ok").unwrap();
 
-        let ov = build_overrides(root, &[], &[]).unwrap();
+        let ov = build_overrides(root, &[], &[], &[]).unwrap();
         let (accepted, skipped_entries) = walk_files_with_skips(root, &ov).unwrap();
 
         let accepted_names: Vec<_> = accepted

crates/kebab-source-fs/tests/include_allowlist.rs

@@ -0,0 +1,111 @@
+//! Integration test: `scope.include` enforces an allow-list.
+//!
+//! Semantics (gitignore convention):
+//!   - `include` is empty Vec → all files pass through (backward-compat).
+//!   - `include` is non-empty → only files matching at least one pattern
+//!     are accepted. `exclude` rules still apply after include.
+//!
+//! Layout (built per-test in a TempDir):
+//!   root/
+//!   ├── a.md
+//!   ├── b.py
+//!   ├── c.png
+//!   └── d.pdf
+
+use std::fs;
+
+use kebab_config::Config;
+use kebab_core::{SourceConnector, SourceScope};
+use kebab_source_fs::FsSourceConnector;
+
+fn cfg_with_root(root: &str) -> Config {
+    let mut c = Config::defaults();
+    c.workspace.root = root.to_string();
+    c.workspace.exclude.clear();
+    // Disable size / generated caps so small test files always pass.
+    c.ingest.code.max_file_bytes = u64::MAX;
+    c.ingest.code.max_file_lines = u32::MAX;
+    c.ingest.code.skip_generated_header = false;
+    c
+}
+
+fn setup_mixed_dir() -> tempfile::TempDir {
+    let dir = tempfile::tempdir().unwrap();
+    let root = dir.path();
+    fs::write(root.join("a.md"), b"md").unwrap();
+    fs::write(root.join("b.py"), b"py").unwrap();
+    fs::write(root.join("c.png"), b"\x89PNG").unwrap();
+    fs::write(root.join("d.pdf"), b"%PDF").unwrap();
+    dir
+}
+
+/// Empty include → all 4 files pass (backward-compat).
+#[test]
+fn include_empty_accepts_all_files() {
+    let dir = setup_mixed_dir();
+    let conn = FsSourceConnector::new(&cfg_with_root(dir.path().to_str().unwrap())).unwrap();
+    let scope = SourceScope {
+        include: vec![],
+        ..SourceScope::default()
+    };
+    let assets = conn.scan(&scope).unwrap();
+    let names: Vec<_> = assets.iter().map(|a| a.workspace_path.0.clone()).collect();
+    assert!(names.contains(&"a.md".to_string()), "a.md missing; got: {names:?}");
+    assert!(names.contains(&"b.py".to_string()), "b.py missing; got: {names:?}");
+    assert!(names.contains(&"c.png".to_string()), "c.png missing; got: {names:?}");
+    assert!(names.contains(&"d.pdf".to_string()), "d.pdf missing; got: {names:?}");
+    assert_eq!(names.len(), 4, "expected exactly 4 files; got: {names:?}");
+}
+
+/// Non-empty include → only md + py come back; png + pdf are excluded.
+#[test]
+fn include_nonempty_is_allowlist() {
+    let dir = setup_mixed_dir();
+    let conn = FsSourceConnector::new(&cfg_with_root(dir.path().to_str().unwrap())).unwrap();
+    let scope = SourceScope {
+        include: vec!["**/*.md".to_string(), "**/*.py".to_string()],
+        ..SourceScope::default()
+    };
+    let assets = conn.scan(&scope).unwrap();
+    let names: Vec<_> = assets.iter().map(|a| a.workspace_path.0.clone()).collect();
+    assert!(names.contains(&"a.md".to_string()), "a.md should be accepted; got: {names:?}");
+    assert!(names.contains(&"b.py".to_string()), "b.py should be accepted; got: {names:?}");
+    assert!(
+        !names.contains(&"c.png".to_string()),
+        "c.png must be rejected by include allowlist; got: {names:?}"
+    );
+    assert!(
+        !names.contains(&"d.pdf".to_string()),
+        "d.pdf must be rejected by include allowlist; got: {names:?}"
+    );
+    assert_eq!(names.len(), 2, "expected exactly 2 files; got: {names:?}");
+}
+
+/// include + exclude are ANDed: a file matching include but also matching
+/// exclude must be rejected.
+#[test]
+fn include_and_exclude_are_anded() {
+    let dir = tempfile::tempdir().unwrap();
+    let root = dir.path();
+    fs::write(root.join("keep.md"), b"keep").unwrap();
+    fs::write(root.join("drop.md"), b"drop").unwrap();
+    fs::write(root.join("other.py"), b"py").unwrap();
+
+    let conn = FsSourceConnector::new(&cfg_with_root(root.to_str().unwrap())).unwrap();
+    let scope = SourceScope {
+        include: vec!["**/*.md".to_string()],
+        exclude: vec!["drop.md".to_string()],
+        ..SourceScope::default()
+    };
+    let assets = conn.scan(&scope).unwrap();
+    let names: Vec<_> = assets.iter().map(|a| a.workspace_path.0.clone()).collect();
+    assert!(names.contains(&"keep.md".to_string()), "keep.md should be accepted; got: {names:?}");
+    assert!(
+        !names.contains(&"drop.md".to_string()),
+        "drop.md should be excluded (matched exclude); got: {names:?}"
+    );
+    assert!(
+        !names.contains(&"other.py".to_string()),
+        "other.py should be excluded (not in include); got: {names:?}"
+    );
+}

crates/kebab-store-sqlite/snapshots/ingest_report.snapshot.json

@@ -56,5 +56,6 @@
   "skipped_kebabignore": 0,
   "skipped_size_exceeded": 0,
   "unchanged": 0,
+  "purged_deleted_files": 0,
   "updated": 1
 }

crates/kebab-store-sqlite/src/documents.rs

@@ -264,6 +264,28 @@ impl kebab_core::DocumentStore for SqliteStore {
         }))
     }
 
+    fn get_asset(
+        &self,
+        id: &kebab_core::AssetId,
+    ) -> Result<Option<kebab_core::RawAsset>> {
+        let conn = self.lock_conn();
+        let result = conn.query_row(
+            r#"SELECT
+                asset_id, source_uri, workspace_path, media_type,
+                byte_len, checksum, storage_kind, storage_path,
+                discovered_at
+            FROM assets
+            WHERE asset_id = ?"#,
+            rusqlite::params![id.0.as_str()],
+            asset_from_row,
+        );
+        match result {
+            Ok(asset) => Ok(Some(asset)),
+            Err(rusqlite::Error::QueryReturnedNoRows) => Ok(None),
+            Err(e) => Err(e.into()),
+        }
+    }
+
     fn get_asset_by_workspace_path(
         &self,
         path: &kebab_core::WorkspacePath,
@@ -286,6 +308,88 @@ impl kebab_core::DocumentStore for SqliteStore {
         }
     }
 
+    fn get_document_by_workspace_path(
+        &self,
+        path: &kebab_core::WorkspacePath,
+    ) -> Result<Option<kebab_core::CanonicalDocument>> {
+        let conn = self.lock_conn();
+        let row: Option<DocumentRow> = conn
+            .query_row(
+                "SELECT
+                    doc_id, asset_id, workspace_path, title, lang,
+                    source_type, trust_level, parser_version,
+                    doc_version, schema_version, metadata_json,
+                    provenance_json, created_at, updated_at,
+                    last_chunker_version, last_embedding_version
+                FROM documents WHERE workspace_path = ?",
+                params![path.0],
+                document_row_from_sql,
+            )
+            .map(Some)
+            .or_else(rows_optional)
+            .map_err(StoreError::from)?;
+        let Some(row) = row else { return Ok(None) };
+
+        let doc_id = kebab_core::DocumentId(row.doc_id.clone());
+        let mut blocks_stmt = conn
+            .prepare(
+                "SELECT payload_json FROM blocks
+                 WHERE doc_id = ? ORDER BY ordinal ASC",
+            )
+            .map_err(StoreError::from)?;
+        let block_rows = blocks_stmt
+            .query_map(params![row.doc_id], |r| {
+                let payload_json: String = r.get(0)?;
+                Ok(payload_json)
+            })
+            .map_err(StoreError::from)?;
+        let mut blocks: Vec<kebab_core::Block> = Vec::new();
+        for block_row in block_rows {
+            let payload_json = block_row.map_err(StoreError::from)?;
+            let block: kebab_core::Block = serde_json::from_str(&payload_json)
+                .context("deserialize block payload_json")?;
+            blocks.push(block);
+        }
+
+        let metadata: kebab_core::Metadata = serde_json::from_str(&row.metadata_json)
+            .context("deserialize metadata_json")?;
+        let provenance: kebab_core::Provenance =
+            serde_json::from_str(&row.provenance_json)
+                .context("deserialize provenance_json")?;
+
+        Ok(Some(kebab_core::CanonicalDocument {
+            doc_id,
+            source_asset_id: kebab_core::AssetId(row.asset_id),
+            workspace_path: kebab_core::WorkspacePath(row.workspace_path),
+            title: row.title.unwrap_or_default(),
+            lang: kebab_core::Lang(row.lang.unwrap_or_default()),
+            blocks,
+            metadata,
+            provenance,
+            parser_version: kebab_core::ParserVersion(row.parser_version),
+            schema_version: row.schema_version as u32,
+            doc_version: row.doc_version as u32,
+            last_chunker_version: row.last_chunker_version.map(kebab_core::ChunkerVersion),
+            last_embedding_version: row.last_embedding_version.map(kebab_core::EmbeddingVersion),
+        }))
+    }
+
+    fn all_workspace_paths(&self) -> Result<Vec<kebab_core::WorkspacePath>> {
+        let conn = self.lock_conn();
+        let mut stmt = conn
+            .prepare("SELECT workspace_path FROM documents")
+            .map_err(StoreError::from)?;
+        let rows = stmt
+            .query_map([], |r| r.get::<_, String>(0))
+            .map_err(StoreError::from)?;
+        let mut out = Vec::new();
+        for row in rows {
+            let path = row.map_err(StoreError::from)?;
+            out.push(kebab_core::WorkspacePath(path));
+        }
+        Ok(out)
+    }
+
     fn list_documents(
         &self,
         filter: &kebab_core::DocFilter,
@@ -550,7 +654,8 @@ fn rows_optional<T>(err: rusqlite::Error) -> rusqlite::Result<Option<T>> {
 
 /// Reconstruct a [`kebab_core::RawAsset`] from one `assets` row.
 /// Row mapper for `RawAsset`. Column names are self-documenting; the
-/// SELECT in [`DocumentStore::get_asset_by_workspace_path`] must include
+/// SELECTs in [`DocumentStore::get_asset`] and
+/// [`DocumentStore::get_asset_by_workspace_path`] must both include
 /// all nine columns by their schema names.
 fn asset_from_row(row: &rusqlite::Row<'_>) -> rusqlite::Result<kebab_core::RawAsset> {
     use std::path::PathBuf;

crates/kebab-store-sqlite/src/filters.rs

@@ -153,6 +153,34 @@ impl SqliteStore {
             }
         }
 
+        // p10-1A-1 fix (dogfood-discovered 2026-05-20): code_lang filter
+        // (IN-list on metadata_json.$.code_lang). Empty Vec = no filter.
+        if !filters.code_lang.is_empty() {
+            let placeholders = std::iter::repeat_n("?", filters.code_lang.len())
+                .collect::<Vec<_>>()
+                .join(",");
+            sql.push_str(&format!(
+                " AND json_extract(d.metadata_json, '$.code_lang') IN ({placeholders})"
+            ));
+            for lang in &filters.code_lang {
+                bind.push(Box::new(lang.clone()));
+            }
+        }
+
+        // p10-1A-1 fix (dogfood-discovered 2026-05-20): repo filter
+        // (IN-list on metadata_json.$.repo). Empty Vec = no filter.
+        if !filters.repo.is_empty() {
+            let placeholders = std::iter::repeat_n("?", filters.repo.len())
+                .collect::<Vec<_>>()
+                .join(",");
+            sql.push_str(&format!(
+                " AND json_extract(d.metadata_json, '$.repo') IN ({placeholders})"
+            ));
+            for repo in &filters.repo {
+                bind.push(Box::new(repo.clone()));
+            }
+        }
+
         // p9-fb-36: ingested_after filter.
         // `documents.updated_at` is RFC3339 TEXT (UTC `Z` per fb-32);
         // lexicographic >= compare is correct — but only when the filter
@@ -408,6 +436,78 @@ mod tests {
             .unwrap();
     }
 
+    /// Variant of `seed_committed_full` that additionally accepts a
+    /// `metadata_json` string so p10-1A-1 filter tests can set
+    /// `metadata.code_lang` / `metadata.repo` without going through the
+    /// full ingest pipeline.
+    #[allow(clippy::too_many_arguments)]
+    fn seed_committed_with_metadata(
+        store: &SqliteStore,
+        chunk_id: &str,
+        doc_id: &str,
+        workspace_path: &str,
+        media_type_json: &str,
+        metadata_json: &str,
+    ) {
+        let asset_id = format!("a{}", &doc_id[..31]);
+        {
+            let conn = store.lock_conn();
+            conn.execute(
+                "INSERT INTO assets (
+                    asset_id, source_uri, workspace_path, media_type, byte_len,
+                    checksum, storage_kind, storage_path, discovered_at
+                 ) VALUES (?, ?, ?, ?, 0, 'deadbeefdeadbeefdeadbeefdeadbeef',
+                           'reference', ?, '1970-01-01T00:00:00Z')",
+                params![
+                    asset_id,
+                    format!("file://{workspace_path}"),
+                    workspace_path,
+                    media_type_json,
+                    workspace_path,
+                ],
+            )
+            .unwrap();
+            conn.execute(
+                "INSERT INTO documents (
+                    doc_id, asset_id, workspace_path, title, lang, source_type,
+                    trust_level, parser_version, doc_version, schema_version,
+                    metadata_json, provenance_json, created_at, updated_at
+                 ) VALUES (?, ?, ?, NULL, 'en', 'code', 'primary', 'v1', 1, 1,
+                           ?, '{}', '1970-01-01T00:00:00Z', '1970-01-01T00:00:00Z')",
+                params![doc_id, asset_id, workspace_path, metadata_json],
+            )
+            .unwrap();
+            conn.execute(
+                "INSERT INTO chunks (
+                    chunk_id, doc_id, text, heading_path_json, section_label,
+                    source_spans_json, token_estimate, chunker_version,
+                    policy_hash, block_ids_json, created_at
+                 ) VALUES (?, ?, 'code snippet', '[]', NULL, '[]', 1, 'v1', 'h', '[]',
+                           '1970-01-01T00:00:00Z')",
+                params![chunk_id, doc_id],
+            )
+            .unwrap();
+        }
+
+        let embed_row = EmbeddingRecordRow {
+            embedding_id: format!("e{}", &chunk_id[..31]),
+            chunk_id: chunk_id.to_string(),
+            model_id: "m".to_string(),
+            model_version: "v1".to_string(),
+            dimensions: 4,
+            lance_table: "t".to_string(),
+            created_at: OffsetDateTime::UNIX_EPOCH,
+        };
+        store
+            .put_embedding_records_pending(std::slice::from_ref(&embed_row))
+            .unwrap();
+        store
+            .mark_embedding_records_committed(std::slice::from_ref(
+                &embed_row.embedding_id,
+            ))
+            .unwrap();
+    }
+
     fn cid(s: &str) -> ChunkId {
         ChunkId(s.to_string())
     }
@@ -671,6 +771,78 @@ mod tests {
         assert_eq!(out, vec![cid(c1)], "doc_id filter must scope to the target doc only");
     }
 
+    // ── p10-1A-1 new filter arms ─────────────────────────────────────────
+
+    #[test]
+    fn filter_chunks_code_lang_keeps_matching_lang() {
+        // c1 = python, c2 = rust, c3 = markdown (no code_lang).
+        // Filter code_lang=["python"] → only c1 survives.
+        let tmp = TempDir::new().unwrap();
+        let store = open_store(&tmp);
+        let c1 = "11111111111111111111111111111111";
+        let c2 = "22222222222222222222222222222222";
+        let c3 = "33333333333333333333333333333333";
+        seed_committed_with_metadata(
+            &store, c1, "d1d1d1d1d1d1d1d1d1d1d1d1d1d1d1d1",
+            "src/main.py", r#""code""#,
+            r#"{"code_lang":"python"}"#,
+        );
+        seed_committed_with_metadata(
+            &store, c2, "d2d2d2d2d2d2d2d2d2d2d2d2d2d2d2d2",
+            "src/lib.rs", r#""code""#,
+            r#"{"code_lang":"rust"}"#,
+        );
+        seed_committed_with_metadata(
+            &store, c3, "d3d3d3d3d3d3d3d3d3d3d3d3d3d3d3d3",
+            "README.md", r#""markdown""#,
+            r#"{}"#,
+        );
+
+        let f = SearchFilters {
+            code_lang: vec!["python".to_string()],
+            ..Default::default()
+        };
+        let out = store
+            .filter_chunks(&[cid(c1), cid(c2), cid(c3)], &f)
+            .unwrap();
+        assert_eq!(out, vec![cid(c1)], "only python chunk should survive code_lang filter");
+    }
+
+    #[test]
+    fn filter_chunks_repo_keeps_matching_repo() {
+        // c1 = repo "httpx", c2 = repo "requests", c3 = no repo.
+        // Filter repo=["httpx"] → only c1 survives.
+        let tmp = TempDir::new().unwrap();
+        let store = open_store(&tmp);
+        let c1 = "11111111111111111111111111111111";
+        let c2 = "22222222222222222222222222222222";
+        let c3 = "33333333333333333333333333333333";
+        seed_committed_with_metadata(
+            &store, c1, "d1d1d1d1d1d1d1d1d1d1d1d1d1d1d1d1",
+            "httpx/client.py", r#""code""#,
+            r#"{"repo":"httpx","code_lang":"python"}"#,
+        );
+        seed_committed_with_metadata(
+            &store, c2, "d2d2d2d2d2d2d2d2d2d2d2d2d2d2d2d2",
+            "requests/api.py", r#""code""#,
+            r#"{"repo":"requests","code_lang":"python"}"#,
+        );
+        seed_committed_with_metadata(
+            &store, c3, "d3d3d3d3d3d3d3d3d3d3d3d3d3d3d3d3",
+            "standalone.py", r#""code""#,
+            r#"{"code_lang":"python"}"#,
+        );
+
+        let f = SearchFilters {
+            repo: vec!["httpx".to_string()],
+            ..Default::default()
+        };
+        let out = store
+            .filter_chunks(&[cid(c1), cid(c2), cid(c3)], &f)
+            .unwrap();
+        assert_eq!(out, vec![cid(c1)], "only httpx chunk should survive repo filter");
+    }
+
     #[test]
     fn filter_chunks_ingested_after_non_utc_offset_compares_as_instant() {
         // Regression test for the non-UTC offset lex-compare bug.

crates/kebab-store-sqlite/src/lib.rs

@@ -35,4 +35,4 @@ pub use error::StoreError;
 pub use eval::{EvalQueryResultRecord, EvalRunRecord, EvalRunRow};
 pub use fts::rebuild_chunks_fts;
 pub use jobs::IngestRunRow;
-pub use store::{CountSummary, NotIndexed, SqliteStore};
+pub use store::{CountSummary, NotIndexed, SqliteStore, purge_deleted_workspace_path};

crates/kebab-store-sqlite/src/store.rs

@@ -540,10 +540,132 @@ pub(crate) fn purge_orphan_at_workspace_path(
     Ok(())
 }
 
+/// Purge all stored data for a document whose on-disk file has been
+/// deleted (as opposed to content-changed, which is handled by
+/// `purge_orphan_at_workspace_path`).
+///
+/// Returns the `chunk_id`s that were associated with the document so
+/// the caller can issue a matching `VectorStore::delete_by_chunk_ids`
+/// on the LanceDB side.
+///
+/// Deletion order:
+/// 1. Collect chunk_ids (before cascade removes them).
+/// 2. DELETE the `documents` row → CASCADE clears `blocks`, `chunks`,
+///    `embedding_records`.
+/// 3. DELETE the `assets` row **only if no other document still
+///    references it** (twin-file protection — `assets` can be shared
+///    across identical-content files via the blake3 PK).
+/// 4. If the asset was `storage_kind = 'copied'`, best-effort delete
+///    the on-disk byte file at `storage_path`.
+///
+/// Returns `Ok(vec![])` when no document exists at `workspace_path`
+/// (idempotent — caller doesn't need to pre-check).
+pub fn purge_deleted_workspace_path(
+    store: &SqliteStore,
+    workspace_path: &kebab_core::WorkspacePath,
+) -> anyhow::Result<Vec<kebab_core::ChunkId>> {
+    let conn = store.lock_conn();
+
+    // Look up the document + its asset_id.
+    let doc_row: Option<(String, String)> = conn
+        .query_row(
+            "SELECT doc_id, asset_id FROM documents WHERE workspace_path = ?",
+            rusqlite::params![workspace_path.0],
+            |r| Ok((r.get(0)?, r.get(1)?)),
+        )
+        .optional()
+        .map_err(StoreError::from)?;
+
+    let Some((doc_id, asset_id)) = doc_row else {
+        return Ok(Vec::new());
+    };
+
+    // 1. Collect chunk_ids before CASCADE removes them.
+    let mut stmt = conn
+        .prepare("SELECT chunk_id FROM chunks WHERE doc_id = ?")
+        .map_err(StoreError::from)?;
+    let rows = stmt
+        .query_map(rusqlite::params![doc_id], |r| r.get::<_, String>(0))
+        .map_err(StoreError::from)?;
+    let chunk_ids: Vec<kebab_core::ChunkId> = rows
+        .map(|r| r.map(kebab_core::ChunkId))
+        .collect::<rusqlite::Result<Vec<_>>>()
+        .map_err(StoreError::from)?;
+    drop(stmt);
+
+    // 2. DELETE the document row (CASCADE clears blocks / chunks /
+    //    embedding_records via the FK constraints in V001).
+    conn.execute(
+        "DELETE FROM documents WHERE doc_id = ?",
+        rusqlite::params![doc_id],
+    )
+    .map_err(StoreError::from)?;
+
+    // 3. Delete the asset row only when no other document still
+    //    references it (twin-file safety: two files with identical
+    //    bytes share a single asset row via the blake3 PK).
+    let remaining_refs: i64 = conn
+        .query_row(
+            "SELECT COUNT(*) FROM documents WHERE asset_id = ?",
+            rusqlite::params![asset_id],
+            |r| r.get(0),
+        )
+        .map_err(StoreError::from)?;
+
+    if remaining_refs == 0 {
+        // 4. Capture storage details before deleting the row.
+        let asset_storage: Option<(String, String)> = conn
+            .query_row(
+                "SELECT storage_kind, storage_path FROM assets WHERE asset_id = ?",
+                rusqlite::params![asset_id],
+                |r| Ok((r.get(0)?, r.get(1)?)),
+            )
+            .optional()
+            .map_err(StoreError::from)?;
+
+        conn.execute(
+            "DELETE FROM assets WHERE asset_id = ?",
+            rusqlite::params![asset_id],
+        )
+        .map_err(StoreError::from)?;
+
+        // 5. Best-effort: remove the on-disk copied asset file.
+        if let Some((storage_kind, storage_path)) = asset_storage {
+            if storage_kind == "copied" {
+                let _ = std::fs::remove_file(&storage_path);
+            }
+        }
+    }
+
+    tracing::debug!(
+        target: "kebab-store-sqlite",
+        workspace_path = %workspace_path.0,
+        doc_id = %doc_id,
+        chunk_count = chunk_ids.len(),
+        "purged deleted-file document from store"
+    );
+
+    Ok(chunk_ids)
+}
+
 /// UPSERT a row into `assets`. Used by both the `put_asset_with_bytes`
 /// path (which has bytes + computed `storage_kind/path`) and the
 /// `DocumentStore::put_asset` path (which only has the `RawAsset` and
 /// reads `storage_kind/path` from `asset.stored`).
+///
+/// **`assets.workspace_path` is "last-registered path" semantics for
+/// twin files** (two source files with identical content share one
+/// `assets` row keyed on `asset_id = blake3(content)`). Each ingest
+/// of either twin overwrites `workspace_path` with whichever path was
+/// seen most recently — this is intentional and correct after PR #146
+/// made `try_skip_unchanged` document-centric (uses
+/// `get_document_by_workspace_path`, not `get_asset_by_workspace_path`)
+/// and PR #149 made `reset --orphans-only` document-centric too.
+/// Do NOT "fix" the flip-flop by adding a UNIQUE constraint on
+/// `workspace_path` in the `assets` table — twin de-dup is load-bearing.
+/// When you need media_type for a known document, use the 2-step lookup
+/// `get_document_by_workspace_path` → `doc.source_asset_id` →
+/// `get_asset(asset_id)` so the result is twin-safe.
 pub(crate) fn upsert_asset_row(
     conn: &Connection,
     asset: &kebab_core::RawAsset,
@@ -701,6 +823,39 @@ impl SqliteStore {
         }
         Ok(out)
     }
+
+    /// p10-1A-2 follow-up (dogfooding 2026-05-20): per-repo doc count for
+    /// `schema.v1`.
+    ///
+    /// Reads `metadata_json->'$.repo'`, groups by the value, and skips rows
+    /// where `repo` is NULL (documents without an explicit repo tag).
+    /// Returns `BTreeMap<String, u32>` — key is the repo name as stored in
+    /// frontmatter, value is the doc count.
+    pub fn repo_breakdown(
+        &self,
+    ) -> anyhow::Result<std::collections::BTreeMap<String, u32>> {
+        use anyhow::Context;
+        let conn = self.read_conn();
+        let mut stmt = conn
+            .prepare(
+                "SELECT json_extract(metadata_json, '$.repo') AS rp, COUNT(*) \
+                 FROM documents \
+                 WHERE rp IS NOT NULL \
+                 GROUP BY rp",
+            )
+            .context("prepare repo_breakdown")?;
+        let rows = stmt
+            .query_map([], |r| {
+                Ok((r.get::<_, String>(0)?, r.get::<_, i64>(1)? as u32))
+            })
+            .context("query repo_breakdown")?;
+        let mut out = std::collections::BTreeMap::new();
+        for row in rows {
+            let (k, v) = row.context("read repo_breakdown row")?;
+            out.insert(k, v);
+        }
+        Ok(out)
+    }
 }
 
 /// Apply the design §5 / task-spec pragmas. Called once per connection.
@@ -817,5 +972,79 @@ mod tests {
         // only one key total
         assert_eq!(bd.len(), 1, "expected exactly 1 entry, got: {bd:?}");
     }
+
+    /// p10-1A-2 follow-up: `repo_breakdown` counts docs by
+    /// `metadata_json.repo`.
+    ///
+    /// Inserts:
+    /// - one doc with `repo = "my-repo"` → must appear with count 1
+    /// - one doc with `repo = null`       → must NOT appear (NULL skipped)
+    ///
+    /// Uses a side rusqlite connection that bypasses the `assets` FK via
+    /// `PRAGMA foreign_keys = OFF` so the test is self-contained.
+    #[test]
+    fn repo_breakdown_counts_by_repo() {
+        let (dir, store) = open_fresh_store();
+
+        let db_path = dir.path().join("kebab.sqlite");
+        let conn = rusqlite::Connection::open(&db_path).unwrap();
+        conn.pragma_update(None, "foreign_keys", "OFF").unwrap();
+
+        // Doc 1: doc with repo = "my-repo"
+        conn.execute(
+            "INSERT INTO documents (
+                doc_id, asset_id, workspace_path,
+                source_type, trust_level, parser_version,
+                doc_version, schema_version,
+                metadata_json, provenance_json,
+                created_at, updated_at
+            ) VALUES (
+                'doc-repo-1', 'asset-r1', 'my-repo/README.md',
+                'markdown', 'primary', 'test-v1',
+                1, 1,
+                '{\"repo\":\"my-repo\"}', '{}',
+                '2024-01-01T00:00:00Z', '2024-01-01T00:00:00Z'
+            )",
+            [],
+        )
+        .unwrap();
+
+        // Doc 2: doc with repo absent (null in JSON)
+        conn.execute(
+            "INSERT INTO documents (
+                doc_id, asset_id, workspace_path,
+                source_type, trust_level, parser_version,
+                doc_version, schema_version,
+                metadata_json, provenance_json,
+                created_at, updated_at
+            ) VALUES (
+                'doc-norepo-1', 'asset-r2', 'standalone/notes.md',
+                'markdown', 'primary', 'test-v1',
+                1, 1,
+                '{\"repo\":null}', '{}',
+                '2024-01-01T00:00:00Z', '2024-01-01T00:00:00Z'
+            )",
+            [],
+        )
+        .unwrap();
+
+        drop(conn); // release side connection before querying via store
+
+        let bd = store.repo_breakdown().unwrap();
+
+        // "my-repo" must appear with count 1
+        assert_eq!(
+            bd.get("my-repo"),
+            Some(&1u32),
+            "expected my-repo=1 in repo_breakdown, got: {bd:?}"
+        );
+        // null repo must NOT appear as any key
+        assert!(
+            !bd.contains_key("null"),
+            "null repo must not appear in breakdown, got: {bd:?}"
+        );
+        // only one key total
+        assert_eq!(bd.len(), 1, "expected exactly 1 entry, got: {bd:?}");
+    }
 }
 

crates/kebab-store-sqlite/tests/ingest_report_snapshot.rs

@@ -41,6 +41,7 @@ fn fixture_report() -> IngestReport {
         skipped_generated: 0,
         skipped_size_exceeded: 0,
         skip_examples: kebab_core::SkipExamples::default(),
+        purged_deleted_files: 0,
         items: Some(vec![
             IngestItem {
                 kind: IngestItemKind::New,

docs/ARCHITECTURE.md

@@ -22,7 +22,8 @@ Cargo workspace, 함수 호출 기반 모듈러 모놀리스. UI binary (`kebab-
 | OCR | Ollama vision LM (default `gemma4:e4b`) — `OcrEngine` trait 으로 Tesseract / Apple Vision 등 future swap (HOTFIXES P6-2) |
 | Image caption | Ollama vision LM, runtime gate `image.caption.enabled` (default OFF) |
 | PDF parser | `lopdf` per-page 텍스트, `chunker_version = "pdf-page-v1"` 가 PDF 자산에 하드코딩 (HOTFIXES P7-3) |
-| code parser | `tree-sitter` + `tree-sitter-rust` — **parser-side** (`kebab-parse-code`), chunker-side 아님 (design §6.3). `chunker_version = "code-rust-ast-v1"`. `ast_chunk_max_lines = 200` 상수 고정 (HOTFIXES 2026-05-19 — Chunker trait 이 per-medium config 미노출). |
+| code parser | `tree-sitter` + `tree-sitter-rust` / `tree-sitter-python` / `tree-sitter-typescript` / `tree-sitter-javascript` / `tree-sitter-go` / `tree-sitter-java` / `tree-sitter-kotlin-ng` — **parser-side** (`kebab-parse-code`), chunker-side 아님 (design §6.3). chunker versions: Rust = `code-rust-ast-v1`, Python = `code-python-ast-v1`, TypeScript = `code-ts-ast-v1`, JavaScript = `code-js-ast-v1`, Go = `code-go-ast-v1`, Java = `code-java-ast-v1`, Kotlin = `code-kotlin-ast-v1`. `ast_chunk_max_lines = 200` 상수 고정 (HOTFIXES 2026-05-19 — Chunker trait 이 per-medium config 미노출). Kotlin grammar 은 `tree-sitter-kotlin-ng` 사용 — bare `tree-sitter-kotlin` 은 tree-sitter 0.21–0.23 에 고착되어 있어 사용 불가. |
+| 1B symbol path | workspace path → module path: Python = dotted prefix (`kebab_eval.metrics.compute_mrr`), TypeScript/JavaScript = slash-style prefix (`src/Foo.Foo.search`). Rust 1A-2 는 file-scope nesting 만 (workspace prefix 없음, 비일관 수용 — HOTFIXES 2026-05-20). |
 | TUI | Ratatui + crossterm — P9-1 Library 패널, P9-2/3/4 진행 예정 |
 | Desktop | Tauri 2 + `pdfjs-dist` (native PDF render backend 금지) — P9-5 |
 | citation 형식 | URI fragment (`path#L12-L34` / `path#p=12` / `path#xywh=0,0,100,50`, W3C Media Fragments) |
@@ -51,7 +52,7 @@ flowchart TB
         ppdf["kebab-parse-pdf"]
         pimg["kebab-parse-image"]
         paud["kebab-parse-audio<br/>(P8 보류)"]
-        pcode["kebab-parse-code<br/>(P10-1A-2)"]
+        pcode["kebab-parse-code<br/>(P10-1A-2 + P10-1B + P10-1C-Go + P10-1C-JK)"]
         ptypes["kebab-parse-types"]
         norm["kebab-normalize"]
         chunk["kebab-chunk"]
@@ -126,6 +127,8 @@ flowchart TB
 
 UI → store/llm/parse 직접 의존 금지. 모든 user-facing 진입은 `kebab-app` facade 만 통한다 (frozen 설계 §8). `kebab-cli` 가 `--config <path>` flag 를 honor 하려면 `kebab_app::*_with_config(cfg, …)` companion 을 통해 Config 을 명시적으로 thread 하는 패턴 — 자세한 이유는 [tasks/HOTFIXES.md](../tasks/HOTFIXES.md) 의 `--config` 항목.
 
+`kebab-parse-code` 의 외부 tree-sitter grammar crate 의존: P10-1A-2 에서 `tree-sitter-rust` 추가, P10-1B 에서 `tree-sitter-python` / `tree-sitter-typescript` / `tree-sitter-javascript` 추가, P10-1C-Go 에서 `tree-sitter-go` 추가, P10-1C-JK 에서 `tree-sitter-java` / `tree-sitter-kotlin-ng` 추가. 모두 `kebab-parse-code` 에만 격리 (facade 룰 — UI crate / chunker 가 직접 import 금지). Kotlin 은 `tree-sitter-kotlin-ng` 사용 (bare `tree-sitter-kotlin` 은 tree-sitter 0.21–0.23 에 고착 — 사용 불가).
+
 ## 디렉토리 구조
 
 ```text
@@ -162,7 +165,7 @@ kebab/
 │   ├── kebab-source-fs/                               # 워크스페이스 walk + checksum (P1-1)
 │   ├── kebab-parse-md/                                # Markdown frontmatter + blocks (P1-2/3)
 │   ├── kebab-normalize/                               # ParsedBlock → CanonicalDocument (P1-4)
-│   ├── kebab-chunk/                                   # heading-aware + pdf-page-v1 + code-rust-ast-v1 chunker (P1-5, P7-2, P10-1A-2)
+│   ├── kebab-chunk/                                   # heading-aware + pdf-page-v1 + code-rust-ast-v1 + code-python-ast-v1 + code-ts-ast-v1 + code-js-ast-v1 + code-go-ast-v1 + code-java-ast-v1 + code-kotlin-ast-v1 chunker (P1-5, P7-2, P10-1A-2, P10-1B, P10-1C-Go, P10-1C-JK)
 │   ├── kebab-store-sqlite/                            # SQLite + FTS5 (V001/V002/V003) (P1-6, P2-1, P3-3)
 │   ├── kebab-search/                                  # Lexical + Vector + Hybrid retriever (P2-2, P3-4)
 │   ├── kebab-embed/  kebab-embed-local/                  # Embedder trait + fastembed adapter (P3-1, P3-2)
@@ -172,7 +175,7 @@ kebab/
 │   ├── kebab-eval/                                    # golden query runner + metrics (P5-1, P5-2)
 │   ├── kebab-parse-image/                             # ImageExtractor + Ollama OCR + caption (P6)
 │   ├── kebab-parse-pdf/                               # lopdf per-page text extractor (P7-1)
-│   ├── kebab-parse-code/                              # tree-sitter Rust AST extractor (P10-1A-2); chunker lives in kebab-chunk
+│   ├── kebab-parse-code/                              # tree-sitter AST extractors: Rust (P10-1A-2), Python + TypeScript + JavaScript (P10-1B), Go (P10-1C-Go), Java + Kotlin (P10-1C-JK — java.rs + kotlin.rs); chunker lives in kebab-chunk
 │   ├── kebab-app/                                     # facade (P0 시그니처 + P3-5/P6-4/P7-3 본체)
 │   ├── kebab-tui/                                     # Ratatui shell + Library 패널 (P9-1)
 │   ├── kebab-mcp/                                     # stdio MCP server — tools: schema, doctor, search, ask (P9-FB-30)

docs/SMOKE.md

@@ -340,6 +340,88 @@ extra_skip_globs = []          # 사용자 추가 skip 패턴
 - `.rs` 파일은 `SourceType::Note` 로 분류됨 (kebab-core `SourceType::Code` variant 미존재). `--media code` filter 는 정상 동작 — `MediaType::Code("rust")` 로 별도 분류됨. 자세한 내용: `tasks/HOTFIXES.md` (2026-05-19 `SourceType::Code` 항목).
 - `.gitignore` 가 honor 됨 — `target/` / `node_modules/` 등은 built-in 안전망으로 자동 skip.
 
+## P10-1B Python / TypeScript / JavaScript 코드 색인
+
+P10-1A-2 와 동일한 격리 KB 설정으로 Python / TypeScript / JavaScript 3 언어를 검증한다. 설정 블록은 P10-1A-2 와 동일 (`[ingest.code]` 절 포함).
+
+```bash
+# 1) 워크스페이스에 Python / TS / JS 파일 추가 (소규모 샘플로 충분)
+mkdir -p /tmp/kebab-smoke/workspace/sample_code
+# Python 예시
+cat > /tmp/kebab-smoke/workspace/sample_code/metrics.py <<'EOF'
+def compute_mrr(results):
+    """Mean Reciprocal Rank."""
+    total = 0.0
+    for i, hit in enumerate(results, 1):
+        if hit:
+            total += 1.0 / i
+            break
+    return total
+EOF
+# TypeScript 예시
+cat > /tmp/kebab-smoke/workspace/sample_code/searcher.ts <<'EOF'
+export class Searcher {
+  search(query: string): string[] {
+    return [];
+  }
+}
+EOF
+# JavaScript 예시
+cat > /tmp/kebab-smoke/workspace/sample_code/utils.js <<'EOF'
+function formatResult(hit) {
+  return `${hit.score}: ${hit.path}`;
+}
+module.exports = { formatResult };
+EOF
+
+# 2) ingest
+KB ingest
+
+# 3) 언어별 검색 (symbol + module path prefix 확인)
+KB search --mode hybrid "compute_mrr" --code-lang python --json | \
+  jq '{hits: [.hits[] | {symbol: .citation.symbol, lang: .citation.lang}]}'
+KB search --mode hybrid "search" --code-lang typescript --json | \
+  jq '{hits: [.hits[] | {symbol: .citation.symbol, lang: .citation.lang}]}'
+KB search --mode hybrid "formatResult" --code-lang javascript --json | \
+  jq '{hits: [.hits[] | {symbol: .citation.symbol, lang: .citation.lang}]}'
+
+# 4) schema stats 에 3 언어 카운트 확인
+KB --json schema | jq '.stats.code_lang_breakdown'
+# 기대: {"python": N, "typescript": N, "javascript": N, "rust": M, ...}
+```
+
+**Symbol path 컨벤션 (2026-05-20 기준)**:
+
+- **Python**: workspace 경로 → dotted module path prefix. `sample_code/metrics.py` 의 `compute_mrr` → symbol `sample_code.metrics.compute_mrr`.
+- **TypeScript / JavaScript**: workspace 경로 → slash-style module path prefix. `sample_code/searcher.ts` 의 `search` → `sample_code/searcher.Searcher.search`. `.tsx` / `.mjs` / `.cjs` / `.jsx` 도 동일 처리.
+- **Rust** (1A-2): file-scope nesting 만, workspace path prefix 없음 (예: `Foo::double`). Python/TS/JS 와 비일관 — HOTFIXES 2026-05-20 참조.
+
+**알려진 동작**:
+
+- `const foo = () => {...}` 같은 expression-level 함수는 `<top-level>` glue 로 잡힘 (declaration-level 단위만 1B 1차 범위). 자세한 내용: `tasks/HOTFIXES.md` (2026-05-20).
+- `.gitignore` honor — `node_modules/` / `__pycache__/` / `.venv/` 등 built-in 안전망 자동 skip.
+
+## P10-1C-Go Go 코드 색인
+
+P10-1B 와 동일한 격리 KB 설정. `.go` 파일을 워크스페이스에 두고 ingest 하면 `code-go-ast-v1` chunker 가 package 단위 AST 로 처리한다.
+
+```bash
+cat > /tmp/kebab-smoke/workspace/sample_code/hello.go <<'EOF'
+package main
+
+import "fmt"
+
+func Hello(name string) string {
+    return fmt.Sprintf("Hello, %s!", name)
+}
+EOF
+
+KB ingest
+KB search --mode hybrid "Hello" --code-lang go --json | \
+  jq '{hits: [.hits[] | {symbol: .citation.symbol, lang: .citation.lang}]}'
+# 기대: symbol = "main.Hello", lang = "go"
+```
+
 ## 검증 체크리스트
 
 - `kebab doctor` 가 `--config` path 를 honor 하고 그 안의 `storage.data_dir` 를 출력 (XDG default 가 아님).
@@ -371,6 +453,9 @@ rm -rf /tmp/kebab-smoke              # 통째로 정리
 - (P7-3) `config.chunking.chunker_version` 는 markdown 만 represent — PDF 자산은 `pdf-page-v1` 하드코딩. `config.toml` 의 `chunker_version = "md-heading-v1"` 을 봐도 PDF 는 영향 안 받음. HOTFIXES `2026-05-02 P7-3` entry 참조 (P+ chunker registry task 까지 유지).
 - (P7-3) 한 PDF 가 N 페이지면 `kebab ingest` 가 N 개 (또는 그 이상의, 페이지 길면 multi-chunk) 의 chunk 를 한 transaction 안에서 commit. 500 페이지 책 → 500+ chunk 한 번에 → embedding throughput 가 bottleneck. 임베딩 활성 워크스페이스에서 큰 PDF 를 처음 ingest 하면 분-단위 시간 + WAL 크기 증가 가능 — P+ 스케일 hardening task 까지 정상 동작이지만 비용은 측정 가능.
 - (P10-1A-2) `.rs` 파일을 워크스페이스에 두면 `kebab ingest` 결과에 `new` 카운터에 포함. `kebab search --mode hybrid "<함수명>" --code-lang rust --json` 가 `citation.kind = "code"`, `citation.lang = "rust"` (SearchHit top-level `code_lang` 도 동일), `citation.symbol` (함수/타입 이름), `citation.line_start` / `citation.line_end` 를 반환하면 wiring 정상. `kebab schema --json | jq .stats.code_lang_breakdown` 에 `"rust": N` 이 나오면 chunk 가 색인됨.
+- (P10-1B) `.py` / `.ts` / `.tsx` / `.js` / `.mjs` / `.cjs` / `.jsx` 파일을 워크스페이스에 두면 `kebab ingest` 결과에 `new` 카운터에 포함. `--code-lang python` / `--code-lang typescript` / `--code-lang javascript` 검색이 `citation.symbol` 에 module path prefix 를 포함한 결과를 반환하면 wiring 정상. `kebab schema --json | jq .stats.code_lang_breakdown` 에 해당 언어 카운트 등장 확인.
+- (P10-1C-Go) `.go` 파일을 워크스페이스에 두면 `kebab ingest` 가 `code-go-ast-v1` 로 처리. `--code-lang go` 검색이 `citation.symbol` 에 `<package>.<Func>` / `<package>.(*Receiver).<Method>` 형식 결과를 반환하면 wiring 정상. `kebab schema --json | jq .stats.code_lang_breakdown` 에 `"go": N` 등장 확인.
+- (P10-1C-JK) `.java` 파일은 `code-java-ast-v1`, `.kt`/`.kts` 파일은 `code-kotlin-ast-v1` 로 처리. `--code-lang java` / `--code-lang kotlin` 검색이 `citation.symbol` 에 `com.foo.Foo.bar` 형식 결과를 반환하면 wiring 정상. `kebab schema --json | jq .stats.code_lang_breakdown` 에 `"java": N` / `"kotlin": N` 등장 확인.
 - (P7-3 + follow-up) 동일 path 에 byte 가 다른 PDF 를 두 번째 ingest 하면 `purge_vector_orphans_for_workspace_path` 가 옛 chunk_id 를 LanceDB 에서 먼저 삭제, 이어서 `purge_orphan_at_workspace_path` 가 옛 doc / chunks / embedding_records 를 SQLite 에서 sweep. 새 byte 가 새 `doc_id` 로 색인됨. `IngestReport` 에 그 자산만 `new+=1` (다른 자산은 `updated`). 두 store 모두 정합 — 옛 본문 검색 시 옛 chunks 가 더 이상 surface 되지 않음.
 
 ### Embedding upgrade (fb-39b)

docs/superpowers/plans/2026-05-20-p10-1b-py-ts-js-ast-chunkers.md

@@ -0,0 +1,741 @@
+# p10-1B Python + TS/JS AST Chunkers Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development to implement this plan task-by-task.
+
+**Goal:** Activate Python / TypeScript / JavaScript code ingest end-to-end on top of 1A-2's infrastructure — 3 new tree-sitter grammars, 3 new Extractors, 3 new chunkers (`code-{python,ts,js}-ast-v1`), a `module_path_for_*` helper for workspace-path → module-path conversion, and a small app-dispatch generalization. Wire `code_lang` filter / breakdown / Citation::Code surface activate automatically.
+
+**Architecture:** Mirror 1A-2 exactly per language. Each Extractor in `kebab-parse-code/src/{python,typescript,javascript}.rs` calls its tree-sitter grammar and emits one `Block::Code` per top-level AST semantic unit with `SourceSpan::Code { line_start, line_end, symbol, lang }`. Symbol = `module_path` (from workspace_path) `+` per-language join (`.` for Python, `/.../basename.symbol` for TS/JS). Each chunker is a near-duplicate of `code-rust-ast-v1` (1:1 + oversize split). App dispatch becomes `match lang { "rust" | "python" | "typescript" | "javascript" }`.
+
+**Tech Stack:** Rust 2024 workspace, `tree-sitter` 0.26, `tree-sitter-python` / `tree-sitter-typescript` / `tree-sitter-javascript`, existing 1A-2 infrastructure (citation_helper Code arm, backfill, schema breakdown).
+
+**Memory note:** Host was OOM-killed earlier in this branch's history. Prefer `cargo test -p <crate>` and `cargo check -p <crate>`; the only `cargo test --workspace -j 1` call is the Task L full-suite gate. Never run cargo invocations in parallel.
+
+---
+
+## Pre-flight
+
+Branch `feat/p10-1b-py-ts-js` already exists on main (`git checkout feat/p10-1b-py-ts-js`).
+
+- [ ] **Disk hygiene**: `cargo clean`.
+
+Reference files (read before touching the corresponding 1B file):
+- 1A-2 Rust extractor: `crates/kebab-parse-code/src/rust.rs` — the scaffold every per-lang extractor mirrors.
+- 1A-2 Rust chunker: `crates/kebab-chunk/src/code_rust_ast_v1.rs` — the scaffold every per-lang chunker mirrors.
+- 1A-2 app dispatch: `crates/kebab-app/src/lib.rs` `ingest_one_code_asset` (~line 1645).
+- 1A-2 source-fs routing: `crates/kebab-source-fs/src/media.rs:39` (the `"rs" =>` arm).
+- 1A-2 lang dispatch: `crates/kebab-parse-code/src/lang.rs::code_lang_for_path`.
+
+---
+
+## Task A: Workspace deps
+
+**Files:**
+- Modify: `Cargo.toml` (workspace `[workspace.dependencies]`, after the existing `tree-sitter-rust` entry)
+- Modify: `crates/kebab-parse-code/Cargo.toml` (`[dependencies]`)
+
+- [ ] **Step 1**: Resolve versions: `cargo add tree-sitter-python tree-sitter-typescript tree-sitter-javascript -p kebab-parse-code`.
+
+- [ ] **Step 2**: Lift the three resolved versions into `[workspace.dependencies]` in the root `Cargo.toml`, immediately after the `tree-sitter-rust` line. Single-line comment first:
+
+```toml
+# Python / TS / JS grammars for code ingest (kebab-parse-code, p10-1B).
+tree-sitter-python     = "<resolved>"
+tree-sitter-typescript = "<resolved>"
+tree-sitter-javascript = "<resolved>"
+```
+
+Then change the crate's `[dependencies]` entries to `{ workspace = true }` matching the existing `tree-sitter` / `tree-sitter-rust` style.
+
+- [ ] **Step 3**: `cargo build -p kebab-parse-code` → clean (unused deps OK; warnings appear when actually imported in later tasks).
+
+- [ ] **Step 4**: Commit.
+
+```bash
+git add Cargo.toml Cargo.lock crates/kebab-parse-code/Cargo.toml
+git commit -m "build(p10-1b): add tree-sitter-python/-typescript/-javascript workspace deps
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task B: source-fs media routing for `.py`/`.pyi`/`.ts`/`.tsx`/`.js`/`.mjs`/`.cjs`/`.jsx`
+
+**Files:**
+- Modify: `crates/kebab-source-fs/src/media.rs` (add 3 arms next to the existing `"rs"` arm at L39)
+- Test: same file's test module
+
+- [ ] **Step 1 (failing test)**:
+
+```rust
+#[test]
+fn py_ts_js_files_map_to_media_code() {
+    assert_eq!(media_type_for(Path::new("a/b.py")),    MediaType::Code("python".into()));
+    assert_eq!(media_type_for(Path::new("a/b.pyi")),   MediaType::Code("python".into()));
+    assert_eq!(media_type_for(Path::new("a/b.ts")),    MediaType::Code("typescript".into()));
+    assert_eq!(media_type_for(Path::new("a/b.tsx")),   MediaType::Code("typescript".into()));
+    assert_eq!(media_type_for(Path::new("a/b.js")),    MediaType::Code("javascript".into()));
+    assert_eq!(media_type_for(Path::new("a/b.mjs")),   MediaType::Code("javascript".into()));
+    assert_eq!(media_type_for(Path::new("a/b.cjs")),   MediaType::Code("javascript".into()));
+    assert_eq!(media_type_for(Path::new("a/b.jsx")),   MediaType::Code("javascript".into()));
+    // Rust 1A-2 arm still works
+    assert_eq!(media_type_for(Path::new("a/b.rs")),    MediaType::Code("rust".into()));
+}
+```
+
+- [ ] **Step 2**: Run → FAIL.
+
+- [ ] **Step 3**: Add the three arms before the `_ => MediaType::Other(ext)` fallback. Match existing style and order extensions logically (most common first within each language):
+
+```rust
+        // p10-1B: Python / TS / JS AST chunkers active.
+        "py" | "pyi"               => MediaType::Code("python".into()),
+        "ts" | "tsx"               => MediaType::Code("typescript".into()),
+        "js" | "mjs" | "cjs" | "jsx" => MediaType::Code("javascript".into()),
+```
+
+- [ ] **Step 4**: Run → PASS. Then `cargo test -p kebab-source-fs` → no regression.
+
+- [ ] **Step 5**: `cargo clippy -p kebab-source-fs --all-targets -- -D warnings` clean. Commit.
+
+```bash
+git add crates/kebab-source-fs/
+git commit -m "feat(p10-1b): route .py/.pyi/.ts/.tsx/.js/.mjs/.cjs/.jsx to MediaType::Code
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task C: `module_path_for_python` + `module_path_for_tsjs` helpers
+
+**Files:**
+- Modify: `crates/kebab-parse-code/src/lang.rs` (add 2 pub fns + tests)
+- Modify: `crates/kebab-parse-code/src/lib.rs` (re-export the 2 fns)
+
+These convert a `WorkspacePath` into a module-path prefix for symbol formatting. Single source of truth — used by all per-language extractors.
+
+### Rules
+
+**`module_path_for_python(workspace_path: &str) -> String`**:
+1. Strip a leading well-known "source root" prefix from a small allowlist if present (in order): `crates/<name>/src/`, `src/`, `lib/`. (Use a single small `for` loop over the allowlist; stop at first prefix match.) Rationale: avoid noisy `crates.x.src.foo.bar` symbols when the user has a conventional layout, while leaving non-conventional paths untouched.
+2. Strip trailing `.py` or `.pyi` extension. If the basename (after extension strip) is `__init__`, drop it (and the preceding `/`) so `pkg/__init__.py` → `pkg`.
+3. Replace `/` with `.`.
+4. Result is the dotted module prefix. Symbols are joined with `.` (e.g. `module_path + "." + sym`). Empty result (file is at workspace root without prefix) → use empty string → symbol is the unit name alone.
+
+**`module_path_for_tsjs(workspace_path: &str) -> String`**:
+1. Strip extension if it's one of `.ts` / `.tsx` / `.js` / `.jsx` / `.mjs` / `.cjs`.
+2. Do NOT replace `/` (TS/JS convention is path-like). Do NOT strip any source root (TS/JS layouts vary too widely).
+3. Result is the path-style prefix (e.g. `src/search/retriever/Retriever`). Symbols join with `.` (`prefix + "." + sym`, e.g. `src/search/retriever/Retriever.search`).
+
+- [ ] **Step 1 (failing tests)** — add to existing `mod tests` (or create one) in `lang.rs`:
+
+```rust
+#[test]
+fn module_path_for_python_strips_src_roots_and_extensions() {
+    assert_eq!(module_path_for_python("kebab_eval/metrics.py"),       "kebab_eval.metrics");
+    assert_eq!(module_path_for_python("kebab_eval/__init__.py"),      "kebab_eval");
+    assert_eq!(module_path_for_python("src/foo/bar.py"),              "foo.bar");
+    assert_eq!(module_path_for_python("crates/x/src/foo/bar.py"),     "foo.bar");
+    assert_eq!(module_path_for_python("a/b/c.pyi"),                   "a.b.c");
+    assert_eq!(module_path_for_python("standalone.py"),               "standalone");
+    assert_eq!(module_path_for_python("src/__init__.py"),             "");
+}
+
+#[test]
+fn module_path_for_tsjs_keeps_slashes_and_strips_ext() {
+    for ext in ["ts", "tsx", "js", "jsx", "mjs", "cjs"] {
+        let p = format!("src/search/retriever/Retriever.{ext}");
+        assert_eq!(module_path_for_tsjs(&p), "src/search/retriever/Retriever");
+    }
+    assert_eq!(module_path_for_tsjs("foo.ts"),                  "foo");
+    assert_eq!(module_path_for_tsjs("a/b/c.ts"),                "a/b/c");
+    // No `src/` strip — TS layouts vary.
+    assert_eq!(module_path_for_tsjs("packages/x/src/Foo.ts"),  "packages/x/src/Foo");
+}
+```
+
+- [ ] **Step 2**: Run → FAIL (helpers not defined).
+
+- [ ] **Step 3**: Implement both in `lang.rs`. Suggested implementation (refine if a test points out a missed edge case):
+
+```rust
+/// p10-1B: workspace-relative Python file path → dotted module-path prefix.
+/// See plan §Task C for the exact rules.
+pub fn module_path_for_python(workspace_path: &str) -> String {
+    let mut p: &str = workspace_path;
+    // Strip a known source-root prefix. Allowlist + `starts_with` over a
+    // pattern with a glob in the middle would be a pain; treat
+    // `crates/*/src/` by string-walking.
+    if let Some(rest) = p.strip_prefix("crates/") {
+        if let Some(slash) = rest.find('/') {
+            let after = &rest[slash + 1..];
+            if let Some(stripped) = after.strip_prefix("src/") {
+                p = stripped;
+            }
+        }
+    } else if let Some(stripped) = p.strip_prefix("src/") {
+        p = stripped;
+    } else if let Some(stripped) = p.strip_prefix("lib/") {
+        p = stripped;
+    }
+    // Strip extension.
+    let p = p
+        .strip_suffix(".py")
+        .or_else(|| p.strip_suffix(".pyi"))
+        .unwrap_or(p);
+    // __init__ → drop it (and the preceding `/`).
+    let p = if let Some(parent) = p.strip_suffix("/__init__") {
+        parent
+    } else if p == "__init__" {
+        ""
+    } else {
+        p
+    };
+    p.replace('/', ".")
+}
+
+/// p10-1B: workspace-relative TS/JS file path → path-style prefix
+/// (no slash replacement). See plan §Task C.
+pub fn module_path_for_tsjs(workspace_path: &str) -> String {
+    let p = workspace_path;
+    for ext in [".tsx", ".ts", ".jsx", ".mjs", ".cjs", ".js"] {
+        if let Some(stripped) = p.strip_suffix(ext) {
+            return stripped.to_string();
+        }
+    }
+    p.to_string()
+}
+```
+
+- [ ] **Step 4**: Re-export both from `lib.rs` (next to the existing `pub use lang::code_lang_for_path`):
+
+```rust
+pub use lang::{code_lang_for_path, module_path_for_python, module_path_for_tsjs};
+```
+
+- [ ] **Step 5**: Run → PASS. clippy clean.
+
+- [ ] **Step 6**: Commit.
+
+```bash
+git add crates/kebab-parse-code/
+git commit -m "feat(p10-1b): module_path_for_python / _tsjs helpers (workspace path → module prefix)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task D: App dispatch generalization
+
+**Files:**
+- Modify: `crates/kebab-app/src/lib.rs`
+
+Today's `ingest_one_code_asset` (~L1645) hardcodes `RustAstExtractor` + `CodeRustAstV1Chunker`. 1B needs to dispatch by `lang`. Cleanest minimal change: keep the same function signature but take `code_lang: &str` and `match` it internally onto an `Extractor` + `Chunker` pair. Rust path keeps the same observable behavior.
+
+Two equivalent dispatch shapes — pick the one with the smallest diff:
+
+**Shape 1 (recommended — fewest lines changed):** factor extractor invocation + chunker invocation into a small `match code_lang` *inside* `ingest_one_code_asset`. The `parser_version` constant lookup also branches. Everything else (read bytes, ExtractContext, put_*, embed, IngestItem) stays a single non-branched flow.
+
+**Shape 2:** introduce a tiny enum `CodeLangKind { Rust, Python, Typescript, Javascript }` + an `impl CodeLangKind { fn extract(...) -> CanonicalDocument; fn chunk(...) -> Vec<Chunk>; fn parser_version() -> ParserVersion; fn chunker_version() -> ChunkerVersion; }`. More structure, but better insulates the function body.
+
+Use Shape 1 for this task (less risk). A future C/D phase can refactor to Shape 2 if the dispatch grows.
+
+- [ ] **Step 1 (failing test)** — add a Python smoke as the failing test (TS/JS land later in this PR; one failing-then-passing TDD cycle is enough to lock the dispatch contract):
+
+In `crates/kebab-app/tests/code_ingest_smoke.rs` add:
+
+```rust
+#[test]
+fn python_file_ingests_and_searches_as_code_citation() {
+    // Mirror rust_file_ingests_and_searches_as_code_citation exactly,
+    // but write `kebab_eval/metrics.py` (in the temp workspace root) with:
+    //     def compute_mrr(): return 1.0
+    // and assert h.code_lang == Some("python"), citation.lang == Some("python"),
+    // citation.symbol == Some("kebab_eval.metrics.compute_mrr"), parser_version "code-python-v1",
+    // chunker_version "code-python-ast-v1".
+    // ...
+}
+```
+
+(Spec shape ONLY — the actual extractor + chunker land in Tasks F + G. This test compiles but FAILS at runtime until those land. Mark it `#[ignore]` if it would otherwise break TDD ordering — un-`#[ignore]` it in Task G's commit. Alternative: skip this step here and rely on the per-extractor unit tests in Task F + Task G; that is the cleaner TDD ordering. Choose either; document the choice in the commit message.)
+
+- [ ] **Step 2**: Update `ingest_one_asset` dispatch match arm to accept all four code languages with a `lang` capture passed through:
+
+```rust
+        // p10-1A-2 / 1B: code ingest dispatch.
+        MediaType::Code(lang)
+            if matches!(lang.as_str(), "rust" | "python" | "typescript" | "javascript") =>
+        {
+            return ingest_one_code_asset(
+                app, asset, chunk_policy, embedder, vector_store,
+                existing_doc_ids, force_reingest, lang.as_str(),
+            );
+        }
+```
+
+(Keep the trailing `MediaType::Code(_) | MediaType::Audio(_) | MediaType::Other(_)` or-pattern as the Skipped fallback — non-allowlisted code langs route there.)
+
+- [ ] **Step 3**: Update `ingest_one_code_asset` signature to take `code_lang: &str` and dispatch internally. Keep all I/O / persistence / embed code unchanged. Per the Shape-1 recipe:
+  - `let parser_version = match code_lang { "rust" => ParserVersion(kebab_parse_code::RUST_PARSER_VERSION.into()), "python" => ParserVersion(kebab_parse_code::PYTHON_PARSER_VERSION.into()), "typescript" => ParserVersion(kebab_parse_code::TS_PARSER_VERSION.into()), "javascript" => ParserVersion(kebab_parse_code::JS_PARSER_VERSION.into()), _ => unreachable!(), };`
+  - The `try_skip_unchanged` call's chunker_version arg branches the same way (different chunker per lang).
+  - The extract call branches: `match code_lang { "rust" => RustAstExtractor::new().extract(...), "python" => PythonAstExtractor::new().extract(...), ... }`.
+  - The chunk call branches: `match code_lang { "rust" => CodeRustAstV1Chunker.chunk(...), "python" => CodePythonAstV1Chunker.chunk(...), ... }`.
+  - All other lines (purge_vector_orphans / put_asset_with_bytes / put_document / put_blocks / put_chunks / embed branch / IngestItem) unchanged.
+
+At this point Python/TS/JS extractors + chunkers don't exist yet → compile FAILS on the references. Acceptable — Task E/F/G/H/I add them. To stage compile-cleanly: gate the Python/TS/JS arms behind `unimplemented!()` for now (returns an error path) and let Tasks F/G/H/I/J/K replace them. Recommended: leave the dispatch fully written but use `anyhow::bail!("not yet activated in this commit")` for the three non-Rust arms, with a `TODO(p10-1b Task X)` comment per arm. They flip to real calls when each language's extractor + chunker land.
+
+- [ ] **Step 4**: `cargo test -p kebab-app --lib` (lib-only is enough — integration tests for the non-Rust paths land later). Existing Rust path tests must stay green.
+
+- [ ] **Step 5**: clippy clean, commit.
+
+```bash
+git add crates/kebab-app/
+git commit -m "refactor(p10-1b): generalize ingest_one_code_asset for multi-language dispatch
+
+Rust path unchanged (verified by existing code_ingest_smoke tests). Python/TS/JS arms
+bail with TODO; per-lang extractor + chunker land in subsequent tasks.
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task E: Python Extractor (`kebab-parse-code/src/python.rs`)
+
+**Files:**
+- Create: `crates/kebab-parse-code/src/python.rs`
+- Modify: `crates/kebab-parse-code/src/lib.rs` (`pub mod python` + re-exports `PYTHON_PARSER_VERSION` and `PythonAstExtractor`)
+- Create: `crates/kebab-parse-code/tests/fixtures/sample.py`
+
+Scaffold MIRRORS `crates/kebab-parse-code/src/rust.rs` line-for-line (read it first). Only the AST walk + the symbol prefix differ.
+
+### Python AST mapping
+
+tree-sitter-python language: `tree_sitter_python::LANGUAGE` (LanguageFn). Set via `parser.set_language(&tree_sitter_python::LANGUAGE.into())`.
+
+Walk `module` (root) named children. Maintain `mod_path: Vec<String>` — but for Python we DO NOT push class names onto `mod_path` (class members get `Class.method` form via the class arm directly; nested classes recurse with the class name appended).
+
+| node kind | unit | symbol (joined with `.`) |
+|-----------|------|--------------------------|
+| `function_definition` (name field) | 1 | `<module_prefix>.<fn_name>` (or `<fn_name>` if module_prefix empty) |
+| `class_definition` (name) — emit ONE unit for the class definition itself (symbol `<module_prefix>.<ClassName>`), then recurse into its `block` body: each inner `function_definition` → unit with symbol `<module_prefix>.<ClassName>.<method_name>`; nested `class_definition` recurses with parent class prepended. | 1 per class + 1 per method (etc.) | as above |
+| `decorated_definition` | unwrap — process its inner `definition` (either function_definition or class_definition) as if at the same level. `unit_start`'s backward extension over `decorator` siblings folds them into the unit. | n/a | n/a |
+| `import_statement`, `import_from_statement`, `expression_statement`, `assignment`, `global_statement`, `future_import_statement` at module level | glue | `<top-level>` (with `module_prefix` prefix if non-empty: `<module_prefix>.<top-level>`) |
+
+`unit_start` (backward extension) covers `comment` siblings + `decorator` siblings (decorators in tree-sitter-python appear as children of `decorated_definition`, NOT as siblings — so the `unwrap decorated_definition` arm above is what brings them in; `comment` siblings still need backward extension). Adapt `unit_start` for the Python flavor: extend over `comment` siblings only (decorators are already covered by unwrapping `decorated_definition`).
+
+Module-prefix application: at extract time, compute `let mod_prefix = kebab_parse_code::module_path_for_python(&asset.workspace_path.0);`. The walk builds symbols using `mod_prefix` (joined with `.` if non-empty; the bare name if empty). Glue group: if `mod_prefix` non-empty, symbol = `format!("{mod_prefix}.<top-level>")`; else `<top-level>`. `<module>` glue label (file contains only `import`s and no real unit) follows the same prefix rule.
+
+### Scaffold differences from rust.rs
+
+- `pub const PARSER_VERSION: &str = "code-python-v1";`
+- `pub struct PythonAstExtractor;` + `new()`/`Default`.
+- `fn supports(&self, m: &MediaType) -> bool { matches!(m, MediaType::Code(l) if l == "python") }`
+- Agent string `"kb-parse-code"` (unchanged).
+- `metadata.code_lang = Some("python".to_string())`.
+- `repo` / `git_branch` / `git_commit` from `crate::repo::detect_repo` (same as Rust).
+- The AST walk is its own `build_blocks` function — DO NOT generalize across languages in this task (each grammar's node names differ enough that polymorphism hurts more than helps; a future refactor task can extract common helpers if patterns converge).
+
+### Step list (TDD)
+
+- [ ] **Step 1**: Create `tests/fixtures/sample.py`:
+
+```python
+"""sample fixture."""
+import os
+
+ANSWER = 42
+
+@staticmethod
+def free(x):
+    """free fn."""
+    return x + 1
+
+class Foo:
+    """doc."""
+    def double(self, n):
+        return n * 2
+
+    @classmethod
+    def name(cls):
+        return "foo"
+
+class Outer:
+    class Inner:
+        def helper(self):
+            return True
+
+def with_decorator():
+    pass
+```
+
+- [ ] **Step 2 (failing test)** in `python.rs`:
+
+```rust
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/sample.py")).unwrap();
+        // Reuse the test-support helper added in Task 6 of 1A-2 (rust.rs tests):
+        // adjust `fixed_rust_asset` to a generic `fixed_code_asset(workspace_path, code_lang)`
+        // OR inline a per-test asset constructor that matches its kebab-core types.
+        let asset = crate::rust::tests_support::fixed_code_asset(
+            "kebab_eval/metrics.py", "python");
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext { asset: &asset, workspace_root: &root, config: &cfg };
+        PythonAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+    #[test]
+    fn extractor_supports_only_media_code_python() {
+        let e = PythonAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("python".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+    #[test]
+    fn python_units_carry_module_prefixed_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc.blocks.iter().map(|b| match b {
+            Block::Code(c) => match &c.common.source_span {
+                SourceSpan::Code { symbol, lang, .. } => {
+                    assert_eq!(lang.as_deref(), Some("python"));
+                    symbol.clone().unwrap()
+                }
+                _ => panic!("expected SourceSpan::Code"),
+            },
+            other => panic!("expected Block::Code, got {other:?}"),
+        }).collect();
+        syms.sort();
+        // workspace_path `kebab_eval/metrics.py` → mod_prefix `kebab_eval.metrics`
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.free"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Foo"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Foo.double"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Foo.name"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Outer"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Outer.Inner"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.Outer.Inner.helper"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.with_decorator"));
+        assert!(syms.iter().any(|s| s == "kebab_eval.metrics.<top-level>"));  // import + assignment
+    }
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 { assert_eq!(extract_fixture().blocks, a.blocks); }
+    }
+}
+```
+
+(`tests_support::fixed_code_asset` — promote 1A-2's `fixed_rust_asset` to a generic helper that takes the lang string and sets `media_type: MediaType::Code(lang.to_string())`. Move it to a new `pub(crate) mod tests_support` in `rust.rs` so it's reachable from `python.rs::tests`, OR duplicate it inline — pick the smaller diff. Keep the helper `#[cfg(test)]`.)
+
+- [ ] **Step 3**: Run → FAIL (`PythonAstExtractor` undefined).
+
+- [ ] **Step 4**: Implement `python.rs`. Scaffold mirrors `rust.rs`; the AST walk follows the table above. The `mod_path: Vec<String>` for Python tracks **class nesting** (so methods get `Class.method`, nested classes get `Outer.Inner`). `Vec` empty at function-level. Glue grouping mirrors Rust's. Apply `mod_prefix` from `module_path_for_python(&asset.workspace_path.0)` to all unit symbols: `if mod_prefix.is_empty() { sym } else { format!("{mod_prefix}.{sym}") }`. The `<top-level>` / `<module>` label inherits the same prefixing.
+
+- [ ] **Step 5**: Wire into `lib.rs`:
+
+```rust
+pub mod python;
+pub use python::{PARSER_VERSION as PYTHON_PARSER_VERSION, PythonAstExtractor};
+```
+
+- [ ] **Step 6**: `cargo test -p kebab-parse-code python` → all pass.
+
+- [ ] **Step 7**: clippy clean, commit.
+
+```bash
+git add crates/kebab-parse-code/
+git commit -m "feat(p10-1b): tree-sitter-python AST extractor (PythonAstExtractor)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task F: Python chunker (`code-python-ast-v1`)
+
+**Files:**
+- Create: `crates/kebab-chunk/src/code_python_ast_v1.rs`
+- Modify: `crates/kebab-chunk/src/lib.rs` (`mod` + `pub use`)
+
+NEAR-DUPLICATE of `crates/kebab-chunk/src/code_rust_ast_v1.rs`. ONLY differences:
+- `const VERSION_LABEL: &str = "code-python-ast-v1";`
+- struct name `CodePythonAstV1Chunker`
+- The validation message says "code-python-ast-v1 only handles..."
+
+`split_oversize` + `make_chunk` + `AST_CHUNK_MAX_LINES` + `BYTES_PER_TOKEN` + `POLICY_HASH_HEX_LEN` IDENTICAL (these are language-agnostic).
+
+- [ ] **Step 1 (failing tests)**: Copy the entire `#[cfg(test)] mod tests` from `code_rust_ast_v1.rs` and substitute `Rust` → `Python` / `code-rust-ast-v1` → `code-python-ast-v1`. Use the same in-memory `code_doc` helper — it doesn't care about the actual language. Add one extra test specifically asserting the `policy_hash` equals the Rust chunker's (cross-chunker fingerprint identity is a 1A-2 invariant — must hold for new chunkers too).
+
+- [ ] **Step 2**: Run → FAIL.
+
+- [ ] **Step 3**: Copy `code_rust_ast_v1.rs` to `code_python_ast_v1.rs` and apply the substitutions above. Keep the `tree-sitter is intentionally NOT a dependency here` comment (still true).
+
+- [ ] **Step 4**: Wire into `lib.rs`:
+
+```rust
+mod code_python_ast_v1;
+pub use code_python_ast_v1::CodePythonAstV1Chunker;
+```
+
+- [ ] **Step 5**: `cargo test -p kebab-chunk code_python_ast` → pass. Full per-crate suite stays green.
+
+- [ ] **Step 6**: clippy clean, commit.
+
+```bash
+git add crates/kebab-chunk/
+git commit -m "feat(p10-1b): code-python-ast-v1 chunker (1:1 + oversize split)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task G: Activate Python in app dispatch
+
+**Files:**
+- Modify: `crates/kebab-app/src/lib.rs` (replace the Python `bail!` arm with real calls)
+- Modify: `crates/kebab-app/tests/code_ingest_smoke.rs` (un-`#[ignore]` the Python test, OR add it now if you deferred in Task D)
+
+- [ ] **Step 1**: Replace the Python arm's `bail!` with `PythonAstExtractor::new().extract(...)` + `CodePythonAstV1Chunker.chunk(...)` calls (mirror the Rust arm exactly). Set parser_version / chunker_version per Python.
+
+- [ ] **Step 2**: Un-ignore / add `python_file_ingests_and_searches_as_code_citation`. Test asserts the full pipeline produces a `Citation::Code { lang: Some("python"), symbol: Some("kebab_eval.metrics.compute_mrr"), .. }` for a `kebab_eval/metrics.py` written into the temp workspace.
+
+- [ ] **Step 3**: `cargo test -p kebab-app code_ingest_smoke python_file_ingests` → pass. Existing Rust test stays green.
+
+- [ ] **Step 4**: clippy clean, commit.
+
+```bash
+git add crates/kebab-app/
+git commit -m "feat(p10-1b): activate Python in ingest_one_code_asset dispatch
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task H: TypeScript Extractor (`kebab-parse-code/src/typescript.rs`)
+
+**Files:**
+- Create: `crates/kebab-parse-code/src/typescript.rs`
+- Modify: `crates/kebab-parse-code/src/lib.rs`
+- Create: `crates/kebab-parse-code/tests/fixtures/sample.ts` + `sample.tsx`
+
+Scaffold mirrors `rust.rs`/`python.rs`. Grammar selection: `tree_sitter_typescript::LANGUAGE_TYPESCRIPT` for `.ts`, `LANGUAGE_TSX` for `.tsx`. Decide inside `extract` by inspecting `asset.workspace_path.0` extension (a tiny helper local to this module is fine).
+
+### TypeScript AST mapping
+
+| node kind | unit | symbol (joined with `.`) |
+|-----------|------|--------------------------|
+| `function_declaration` (name) | 1 | `<mod>.<fn>` |
+| `class_declaration` (name) — recurse into `class_body`: each `method_definition` (name) → unit `<mod>.<Class>.<method>` | 1 + 1 per method | as above |
+| `interface_declaration` (name), `type_alias_declaration` (name), `enum_declaration` (name) | 1 | `<mod>.<Name>` |
+| `export_statement` wrapping any of the above | unwrap to inner declaration; if the inner is `class_declaration` / `function_declaration` / `interface_declaration` / `type_alias_declaration` / `enum_declaration`, treat as that arm. If `export_statement` itself contains a default (i.e., `export default function () {...}` with no name field), emit unit symbol `<mod>.default`. | unwrapped as above, OR `<mod>.default` for nameless default |
+| `lexical_declaration` / `variable_declaration` at top level (`const`/`let`/`var`) | glue | `<top-level>` (prefixed) |
+| `import_statement`, `export_statement` of bare values | glue | as above |
+
+`mod_path` for TS is empty (TS modules are file-level, not nested class/namespace at the symbol level — interfaces/types DO live in module scope but their names are unit-level, not parent context). Skip TS `namespace` / `module` declarations: emit them as glue for 1B (the explicit-namespace case is rare in modern TS; documented in 1B Risks).
+
+Module prefix: `mod_prefix = module_path_for_tsjs(&asset.workspace_path.0)`. Join with `.` for symbol.
+
+### Steps
+
+- [ ] **Step 1 (fixtures)**:
+
+```typescript
+// sample.ts
+import { x } from "./other";
+const ANSWER = 42;
+export interface Greet { hello(): string; }
+export type Maybe<T> = T | null;
+export function add(a: number, b: number): number { return a + b; }
+export class Retriever {
+    search(q: string): string[] { return []; }
+    static create(): Retriever { return new Retriever(); }
+}
+export default function () { return 1; }
+```
+
+```tsx
+// sample.tsx
+import React from "react";
+export function Hello({ name }: { name: string }) { return <span>{name}</span>; }
+export const App = () => <Hello name="x" />;  // arrow fn assigned → glue in 1B
+```
+
+- [ ] **Step 2 (failing tests)**: 2 fixture-based tests asserting per-fixture symbols. Asserted symbols (sample.ts):
+  - `src/sample.add` (if workspace_path is `src/sample.ts`)
+  - `src/sample.Greet`, `src/sample.Maybe`, `src/sample.Retriever`, `src/sample.Retriever.search`, `src/sample.Retriever.create`, `src/sample.default`, `src/sample.<top-level>`.
+- For sample.tsx (workspace_path `src/sample.tsx`): `src/sample.Hello`, `src/sample.<top-level>` (App arrow fn rolled into glue).
+- Also: `extractor_supports_only_media_code_typescript`, `deterministic_across_runs`.
+
+- [ ] **Step 3**: Run → FAIL.
+
+- [ ] **Step 4**: Implement `typescript.rs` mirroring `rust.rs` scaffold. Grammar selection by file extension. AST walk per the table above. Module prefix application same shape as Python (prefix joined with `.`).
+
+- [ ] **Step 5**: Wire into `lib.rs`:
+
+```rust
+pub mod typescript;
+pub use typescript::{PARSER_VERSION as TS_PARSER_VERSION, TypescriptAstExtractor};
+```
+
+- [ ] **Step 6**: Tests pass, clippy clean, commit.
+
+---
+
+## Task I: TS chunker (`code-ts-ast-v1`)
+
+Pattern identical to Task F — duplicate `code_rust_ast_v1.rs` with substitutions (`VERSION_LABEL = "code-ts-ast-v1"`, struct `CodeTsAstV1Chunker`, error message). Test module copies the Rust chunker tests with name substitutions + adds `policy_hash_matches_md_heading_v1`.
+
+Commit:
+
+```
+feat(p10-1b): code-ts-ast-v1 chunker (1:1 + oversize split)
+```
+
+---
+
+## Task J: Activate TypeScript in app dispatch
+
+Mirror Task G. Replace TS `bail!` arm with real calls. Add `typescript_file_ingests_and_searches_as_code_citation` integration test using a `src/Foo.ts` fixture.
+
+Commit:
+
+```
+feat(p10-1b): activate TypeScript in ingest_one_code_asset dispatch
+```
+
+---
+
+## Task K: JavaScript Extractor (`javascript.rs`)
+
+Mirror Task H. tree-sitter-javascript single LanguageFn. AST mapping similar to TS but without `interface_declaration` / `type_alias_declaration` / `enum_declaration`. Module prefix via `module_path_for_tsjs`.
+
+Test fixture `sample.js`:
+
+```javascript
+// sample.js
+import { x } from "./other";
+const ANSWER = 42;
+export function add(a, b) { return a + b; }
+export class Retriever {
+    search(q) { return []; }
+    static create() { return new Retriever(); }
+}
+export default function () { return 1; }
+```
+
+Asserted symbols: `src/sample.add`, `src/sample.Retriever`, `src/sample.Retriever.search`, `src/sample.Retriever.create`, `src/sample.default`, `src/sample.<top-level>`.
+
+Wire into `lib.rs`:
+
+```rust
+pub mod javascript;
+pub use javascript::{PARSER_VERSION as JS_PARSER_VERSION, JavascriptAstExtractor};
+```
+
+Commits:
+
+```
+feat(p10-1b): tree-sitter-javascript AST extractor (JavascriptAstExtractor)
+```
+
+---
+
+## Task L: JS chunker (`code-js-ast-v1`) + Activate JS in app dispatch
+
+Combine Task F + Task G shape for JS in a single commit (less ceremony than splitting since the diffs are tiny):
+
+- Chunker: duplicate-with-substitution from `code_rust_ast_v1.rs`. `VERSION_LABEL = "code-js-ast-v1"`, struct `CodeJsAstV1Chunker`.
+- App dispatch: replace JS `bail!` with real calls.
+- Integration test: `javascript_file_ingests_and_searches_as_code_citation`.
+
+Commit:
+
+```
+feat(p10-1b): code-js-ast-v1 chunker + activate JS in app dispatch
+```
+
+---
+
+## Task M: Snapshots + full-suite gate + manual SMOKE
+
+**Files:**
+- Create: `crates/kebab-chunk/tests/code_python_ast_snapshot.rs` + fixture `tests/fixtures/code-sample.py` + baseline `code-sample.chunks.snapshot.json`
+- Create: same for TS (`code_ts_ast_snapshot.rs` + fixture `.ts` + baseline)
+- Create: same for JS (`code_js_ast_snapshot.rs` + fixture `.js` + baseline)
+
+Mirror `crates/kebab-chunk/tests/code_rust_ast_snapshot.rs` exactly for each language. Build the `CanonicalDocument` IN-MEMORY (no `kebab-parse-code` dep crossing the chunk boundary).
+
+- [ ] **Step 1**: Add the 3 snapshot tests. Generate baselines: `UPDATE_SNAPSHOTS=1 cargo test -p kebab-chunk code_{python,ts,js}_ast_snapshot`. Re-run without env var → PASS.
+
+- [ ] **Step 2**: Full-suite gate (memory-conscious):
+  - `cargo clippy --workspace --all-targets -- -D warnings` (one invocation, no parallel).
+  - `cargo test --workspace --no-fail-fast -j 1` (the `-j 1` is mandatory). If the pre-existing `runner_lexical_is_deterministic_per_query_payload` flake reappears (unlikely — was fixed in PR #141 on main and merged before 1B branch was cut), re-run that single test once.
+
+- [ ] **Step 3**: Manual SMOKE (mirror `docs/SMOKE.md` P10-1A-2 flow for each language):
+
+```bash
+cargo build --release
+rm -rf /tmp/kebab-1bsmoke && mkdir -p /tmp/kebab-1bsmoke/ws/{kebab_eval,src}
+echo 'def compute_mrr(): return 1.0' > /tmp/kebab-1bsmoke/ws/kebab_eval/metrics.py
+echo 'export function add(a,b){return a+b;}' > /tmp/kebab-1bsmoke/ws/src/foo.ts
+echo 'export function sub(a,b){return a-b;}' > /tmp/kebab-1bsmoke/ws/src/bar.js
+# (match isolated config block format from docs/SMOKE.md)
+./target/release/kebab --config /tmp/kebab-1bsmoke/config.toml ingest --json | jq '.items[].parser_version' | sort -u
+./target/release/kebab --config /tmp/kebab-1bsmoke/config.toml search "compute_mrr" --code-lang python --json | jq '.hits[0]'
+./target/release/kebab --config /tmp/kebab-1bsmoke/config.toml schema --json | jq '.stats.code_lang_breakdown'
+```
+
+Expected: parser_versions include `code-python-v1`, `code-ts-v1`, `code-js-v1`. Search returns `Citation::Code { lang: "python", symbol: "kebab_eval.metrics.compute_mrr" }`. `code_lang_breakdown` includes all four langs (rust may be 0 unless you also added a .rs).
+
+- [ ] **Step 4**: Commit (snapshot files + any harness tweaks).
+
+```bash
+git add crates/kebab-chunk/tests/
+git commit -m "test(p10-1b): per-language chunker snapshots + full-suite gate"
+```
+
+---
+
+## Task N: Docs + HOTFIXES + version bump
+
+- README: 지원 형식 / 명령 table row adds Python / TypeScript / JavaScript next to Rust. Mermaid stays unchanged (no new external surface crosses the diagram).
+- HANDOFF: P10 row notes 1B merged (3 langs active). Add a one-line entry under 머지 후 결정 cross-linking the HOTFIXES entries.
+- ARCHITECTURE: dependency-graph edge `pcode → core` already present. The new tree-sitter-{python,typescript,javascript} edges to `pcode` add to the description text. Locked-in decisions table: add "1B symbol path: workspace path → module path (Python dotted, TS/JS slash-style); Rust 1A keeps file-scope nesting only — HOTFIXES 2026-05-20".
+- SMOKE: add 1B section mirroring the 1A-2 P10 section structure (config block, ingest / search / schema verification commands) for Python and TS/JS. Compact — one shared section for all three.
+- tasks/INDEX + tasks/p10/INDEX: flip 1B row 🟡→🟢 (on PR open; ✅ on merge).
+- tasks/HOTFIXES.md: TWO dated 2026-05-20 entries:
+  1. **Rust 1A-2 symbol path is file-scope-only; 1B+ uses workspace path → module prefix**. Cross-link to design §3.4. Acceptable inconsistency for now (cost of 1A retrofit = chunker_version bump + reindex for every existing Rust corpus). User-requested retrofit triggers a separate task.
+  2. **Expression-level functions (arrow fn / function expression assigned to const) NOT emitted as separate units in 1B 1차**. They fold into the `<top-level>` glue. Documented limit; future phase may add `lexical_declaration` → inner-expression unwrap.
+  Cross-link both in `tasks/p10/p10-1b-py-ts-js-ast-chunkers.md` Risks/notes.
+- `docs/superpowers/specs/2026-04-27-kebab-final-form-design.md` §10.1: add a one-liner — "p10-1B 활성화 (Python / TypeScript / JavaScript)".
+- `Cargo.toml`: workspace version `0.7.0 → 0.8.0`. `cargo build --release` refreshes Cargo.lock.
+- One commit:
+
+```bash
+git add -A
+git commit -m "docs(p10-1b): README/HANDOFF/ARCHITECTURE/SMOKE/INDEX + HOTFIXES; chore: bump version 0.7.0 → 0.8.0
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Finalize
+
+- `gitea-pr` open the PR (gitea-ops skill) — title `feat(p10-1B): Python + TS/JS AST chunkers — tree-sitter-{python,typescript,javascript} 코드 색인 활성화`.
+- **Review loop mode** (fixed per workflow memory) until APPROVE → merge → main pull → branch cleanup → `cargo clean` → `gitea-release v0.8.0`.
+
+---
+
+## Self-review checklist (filled by plan author)
+
+- **Spec coverage**: every row of design §1B has a task; §3.4 symbol path covered by Task C + per-language extractors + integration tests; §6.1/§6.2 module structure covered by Tasks E/F/H/I/K/L; §9.1 Tier-1 + oversize fallback inherited from 1A-2 chunker pattern (Tasks F/I/L); §3.5 code_lang already in 1A-2 helper, extended in Task B routing; §5 dispatch covered by Task D; cascade rule (versioning §9) — chunker versions are per-language, fixture snapshots lock behavior.
+- **No placeholders**: all novel logic (module_path helpers, app dispatch generalization, Python AST walk rules) given concretely with full code or exact deltas vs 1A-2. The per-language chunkers are explicit "duplicate code_rust_ast_v1.rs with substitution X/Y/Z" — concrete and verifiable, not vague.
+- **Type consistency**: parser_version constants (`code-{rust,python,ts,js}-v1`) and chunker_version labels (`code-{rust,python,ts,js}-ast-v1`) used consistently across Tasks D/E/F/G/H/I/J/K/L. `module_path_for_python` / `module_path_for_tsjs` referenced consistently as the source of truth for prefixing.

docs/superpowers/plans/2026-05-20-p10-1c-go-ast-chunker.md

@@ -0,0 +1,540 @@
+# p10-1C-Go Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development to implement this plan task-by-task.
+
+**Goal:** Activate Go code ingest end-to-end on top of 1A-2 (Rust) + 1B (Python/TS/JS) infrastructure. Add `tree-sitter-go` grammar + `GoAstExtractor` + `code-go-ast-v1` chunker + media routing + app dispatch arm.
+
+**Architecture:** Mirror 1A-2 / 1B exactly. `kebab-parse-code/src/go.rs` walks tree-sitter-go parse tree; emits one `Block::Code` per top-level AST semantic unit with `SourceSpan::Code { symbol, lang: Some("go") }`. Symbol prefix = **source-extracted package name** (from `package_clause` AST node — design §3.4 Go row). `kebab-chunk/src/code_go_ast_v1.rs` is a near-duplicate of `code-rust-ast-v1`. App dispatch's `ingest_one_code_asset` (PR #142 generalized 4-arm match) gets a 5th arm.
+
+**Tech Stack:** Rust 2024 workspace, `tree-sitter` 0.26 (already in workspace), `tree-sitter-go` (NEW), 1A-2/1B infrastructure unchanged.
+
+**Memory note:** Host has been OOM-killed previously. Use `cargo test -p <crate>` and `cargo check -p <crate>` only. ONE full-suite invocation reserved for Task G gate.
+
+---
+
+## Pre-flight
+
+Branch `feat/p10-1c-go` already exists.
+
+- [ ] **Disk hygiene**: `cargo clean` if previous artifacts are bloated. Skip if disk is comfortable (`df -h /`).
+
+Reference files:
+- 1A-2 Rust extractor: `crates/kebab-parse-code/src/rust.rs` — closest single-language scaffold template.
+- 1B Python extractor (closest analog for "class-nesting recursion" — Go doesn't have classes but has package as the single prefix): `crates/kebab-parse-code/src/python.rs`.
+- 1A-2 chunker scaffold: `crates/kebab-chunk/src/code_rust_ast_v1.rs`.
+- 1B dispatch generalization: `crates/kebab-app/src/lib.rs::ingest_one_code_asset` (~L1645, 4-arm match).
+- 1A-2 source-fs routing: `crates/kebab-source-fs/src/media.rs` `"rs" =>` arm.
+
+---
+
+## Task A: Workspace dep `tree-sitter-go`
+
+**Files:**
+- Modify: `Cargo.toml` (workspace `[workspace.dependencies]`, after `tree-sitter-javascript` line)
+- Modify: `crates/kebab-parse-code/Cargo.toml`
+
+- [ ] **Step 1**: `cargo add tree-sitter-go -p kebab-parse-code` to resolve version.
+
+- [ ] **Step 2**: Lift the resolved version into `[workspace.dependencies]` after `tree-sitter-javascript`:
+
+```toml
+# Go grammar for code ingest (kebab-parse-code, p10-1C).
+tree-sitter-go         = "<resolved>"
+```
+
+Switch the crate's entry to `{ workspace = true }` matching existing tree-sitter-* style.
+
+- [ ] **Step 3**: `cargo build -p kebab-parse-code` → clean. Unused dep warning is fine.
+
+- [ ] **Step 4**: Commit:
+
+```bash
+git add Cargo.toml Cargo.lock crates/kebab-parse-code/Cargo.toml
+git commit -m "build(p10-1c-go): add tree-sitter-go workspace dep
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task B: source-fs media routing `.go` → `MediaType::Code("go")`
+
+**Files:**
+- Modify: `crates/kebab-source-fs/src/media.rs` (add arm after the existing JS arm at ~L44)
+- Test: same file's test module
+
+- [ ] **Step 1 (failing test)** — add to existing tests near `py_ts_js_files_map_to_media_code`:
+
+```rust
+#[test]
+fn go_files_map_to_media_code_go() {
+    assert_eq!(media_type_for(Path::new("a/b.go")), MediaType::Code("go".into()));
+}
+```
+
+- [ ] **Step 2**: Run → FAIL.
+
+- [ ] **Step 3**: Add the arm before the catch-all `_ => MediaType::Other(ext)`:
+
+```rust
+        // p10-1C-Go: Go ingest activated.
+        "go" => MediaType::Code("go".into()),
+```
+
+- [ ] **Step 4**: Run → PASS. `cargo test -p kebab-source-fs` → no regression.
+
+- [ ] **Step 5**: clippy clean, commit:
+
+```bash
+git add crates/kebab-source-fs/
+git commit -m "feat(p10-1c-go): route .go to MediaType::Code(go)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task C: App dispatch allowlist + bail arm for "go"
+
+**Files:**
+- Modify: `crates/kebab-app/src/lib.rs` (dispatch match guard + 4 internal match arms in `ingest_one_code_asset`)
+
+- [ ] **Step 1**: Find the `MediaType::Code(lang) if matches!(lang.as_str(), "rust" | "python" | "typescript" | "javascript")` arm (~L953). Add `"go"` to the allowlist:
+
+```rust
+        MediaType::Code(lang)
+            if matches!(lang.as_str(), "rust" | "python" | "typescript" | "javascript" | "go") =>
+        {
+```
+
+- [ ] **Step 2**: In `ingest_one_code_asset`'s 4 `match code_lang` blocks (parser_version, chunker_version, extract, chunk), add a "go" arm that `bail!()`s for now (extractor + chunker land in Task D/E). Mirror the Python/TS/JS bail-then-activate pattern:
+
+```rust
+let parser_version = match code_lang {
+    // ... existing arms ...
+    "go" => anyhow::bail!("go ingest not yet wired (p10-1c-go Task F)"),
+    other => anyhow::bail!("unsupported code_lang: {other}"),
+};
+// similar for chunker_version / extract / chunk matches
+```
+
+- [ ] **Step 3**: `cargo test -p kebab-app --lib` → existing 52 lib tests stay green. `cargo test -p kebab-app --test code_ingest_smoke` → 6 stay green (Rust path unaffected).
+
+- [ ] **Step 4**: clippy clean, commit:
+
+```bash
+git add crates/kebab-app/
+git commit -m "refactor(p10-1c-go): add go to ingest dispatch allowlist (bail until Task F)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task D: `GoAstExtractor` (`kebab-parse-code/src/go.rs`)
+
+**Files:**
+- Create: `crates/kebab-parse-code/src/go.rs`
+- Modify: `crates/kebab-parse-code/src/lib.rs` (`pub mod go;` + re-exports `GO_PARSER_VERSION`, `GoAstExtractor`)
+- Create: `crates/kebab-parse-code/tests/fixtures/sample.go`
+
+Scaffold mirrors `crates/kebab-parse-code/src/rust.rs` line-for-line for the `CanonicalDocument` skeleton (Extractor trait impl, `id_for_doc`, ProvenanceEvent, final `CanonicalDocument` literal). The novel parts:
+
+### Constants
+
+```rust
+pub const PARSER_VERSION: &str = "code-go-v1";
+
+pub struct GoAstExtractor;
+// new() + Default
+// supports: matches!(m, MediaType::Code(l) if l == "go")
+// agent = "kb-parse-code"
+// metadata.code_lang = Some("go")
+// SourceType::Note (no SourceType::Code variant)
+// repo/git_branch/git_commit via detect_repo
+```
+
+### Package extraction
+
+Unlike 1B's path-based `module_path_for_python` / `_for_tsjs`, the Go package prefix comes from the **source code's `package` declaration** (design §3.4). tree-sitter-go's grammar:
+
+- Root: `source_file`
+- First named child is typically `package_clause` → contains `package_identifier` child whose text is the package name.
+
+Helper (local to `go.rs`):
+
+```rust
+/// Returns the package name from a tree-sitter-go `source_file`, or
+/// `None` if the file has no `package_clause` (invalid Go in practice,
+/// but be defensive).
+fn extract_package(root: tree_sitter::Node, src: &str) -> Option<String> {
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        if child.kind() == "package_clause" {
+            // `package_clause` has a `package_identifier` named child.
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                if sub.kind() == "package_identifier" {
+                    return Some(src[sub.start_byte()..sub.end_byte()].to_string());
+                }
+            }
+        }
+    }
+    None
+}
+```
+
+### Semantic-unit rules
+
+| node kind | unit | symbol |
+|-----------|------|--------|
+| `function_declaration` (name field) | 1 | `<pkg>.<fn_name>` |
+| `method_declaration` | 1 | `<pkg>.(<TypeText>).<MethodName>` where `<TypeText>` includes a leading `*` if the receiver is `pointer_type`. Examples: `chunk.(*MdHeadingV1Chunker).ChunkDoc`, `chunk.(Foo).Bar`. |
+| `type_declaration` (struct / interface / type alias) | 1 per inner `type_spec` | `<pkg>.<TypeName>` |
+| `const_declaration`, `var_declaration`, `import_declaration` (single or block) | glue | `<pkg>.<top-level>` (or `<pkg>.<package>` if file has ZERO real units AND glue is import-only — same `<module>` post-pass pattern as 1B Python, renamed to `<package>` to avoid colliding with Go's `package` keyword? — actually use `<module>` per design §3.4 — see "module / namespace 만 있고 symbol 없는 경우" line) |
+
+`unit_start` walks `comment` siblings (same as 1B). Go doesn't have separate attribute / decorator nodes.
+
+Method receiver pointer detection:
+
+```rust
+// In the method_declaration arm:
+let receiver = child.child_by_field_name("receiver");  // parameter_list
+let receiver_type_text = receiver.and_then(|r| {
+    let mut cw = r.walk();
+    for p in r.named_children(&mut cw) {
+        if p.kind() == "parameter_declaration" {
+            // type field is either type_identifier (value) or pointer_type (ptr)
+            if let Some(ty) = p.child_by_field_name("type") {
+                let s = &src[ty.start_byte()..ty.end_byte()];
+                return Some(s.to_string());  // includes leading "*" if pointer_type
+            }
+        }
+    }
+    None
+});
+// Format: "(*Foo)" or "(Foo)" — wrap in parens, preserve leading "*" if any.
+let owner = receiver_type_text
+    .map(|t| format!("({t})"))
+    .unwrap_or_else(|| "()".to_string());
+let method_name = name_text(&child, src);
+// symbol = format!("{pkg}.{owner}.{method_name}")
+```
+
+Read tree-sitter-go's grammar.json or node-types.json (in the registry source) if any field name above differs in the resolved crate version.
+
+### Fixture `tests/fixtures/sample.go`:
+
+```go
+// sample.go
+package chunk
+
+import (
+	"fmt"
+	"strings"
+)
+
+const Version = "v1"
+
+type MdHeadingV1Chunker struct {
+	Name string
+}
+
+// ChunkDoc returns a stub list of strings.
+func (m *MdHeadingV1Chunker) ChunkDoc(input string) []string {
+	return []string{m.Name}
+}
+
+func (m MdHeadingV1Chunker) Name2() string {
+	return m.Name
+}
+
+type Stringer interface {
+	String() string
+}
+
+func Free(x int) int {
+	return x + 1
+}
+
+func init() {
+	fmt.Println(strings.ToUpper("init"))
+}
+```
+
+### Test module
+
+Mirror Python's test shape (use `crate::rust::tests_support::fixed_code_asset` from 1B):
+
+```rust
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(
+            concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/sample.go"),
+        ).unwrap();
+        let asset = crate::rust::tests_support::fixed_code_asset(
+            "crates/x/src/sample.go", "go",
+        );
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext { asset: &asset, workspace_root: &root, config: &cfg };
+        GoAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_go() {
+        let e = GoAstExtractor::new();
+        assert!(e.supports(&MediaType::Code("go".into())));
+        assert!(!e.supports(&MediaType::Code("rust".into())));
+        assert!(!e.supports(&MediaType::Markdown));
+    }
+
+    #[test]
+    fn go_units_match_design_3_4_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc.blocks.iter().filter_map(|b| match b {
+            Block::Code(c) => match &c.common.source_span {
+                SourceSpan::Code { symbol, lang, .. } => {
+                    assert_eq!(lang.as_deref(), Some("go"));
+                    symbol.clone()
+                }
+                _ => None,
+            },
+            _ => None,
+        }).collect();
+        syms.sort();
+        assert!(syms.iter().any(|s| s == "chunk.Free"), "got {syms:?}");
+        assert!(syms.iter().any(|s| s == "chunk.init"));
+        assert!(syms.iter().any(|s| s == "chunk.MdHeadingV1Chunker"));
+        assert!(syms.iter().any(|s| s == "chunk.(*MdHeadingV1Chunker).ChunkDoc"));
+        assert!(syms.iter().any(|s| s == "chunk.(MdHeadingV1Chunker).Name2"));
+        assert!(syms.iter().any(|s| s == "chunk.Stringer"));
+        assert!(syms.iter().any(|s| s == "chunk.<top-level>"));  // import + const grouped
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 { assert_eq!(extract_fixture().blocks, a.blocks); }
+    }
+}
+```
+
+### Step list
+
+- [ ] Step 1: create fixture + test module.
+- [ ] Step 2: run → FAIL (`GoAstExtractor` undefined).
+- [ ] Step 3: implement `go.rs`. Scaffold mirrors `python.rs` (Extractor impl + extract scaffold + `build_blocks` returning blocks). `build_blocks` does: extract_package → walk root's named children → branch per node kind per the table above → emit `Block::Code` with `SourceSpan::Code { symbol, lang: Some("go") }`. Use the same `flush_glue` / glue grouping / `<top-level>` vs `<module>` post-pass as Python (rename to `<package>` if user prefers, but spec §3.4 says `<module>` so keep that name for cross-language consistency).
+- [ ] Step 4: wire into `lib.rs`:
+
+```rust
+pub mod go;
+pub use go::{PARSER_VERSION as GO_PARSER_VERSION, GoAstExtractor};
+```
+
+- [ ] Step 5: `cargo test -p kebab-parse-code` → all pass (Rust/Python/TS/JS + new Go). `cargo clippy -p kebab-parse-code --all-targets -- -D warnings` clean.
+- [ ] Step 6: commit:
+
+```bash
+git add crates/kebab-parse-code/
+git commit -m "feat(p10-1c-go): tree-sitter-go AST extractor (GoAstExtractor)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task E: `code-go-ast-v1` chunker
+
+**Files:**
+- Create: `crates/kebab-chunk/src/code_go_ast_v1.rs`
+- Modify: `crates/kebab-chunk/src/lib.rs`
+
+Identical pattern to PR #142 Task I (TS) / Task L (JS) — near-duplicate of `code_rust_ast_v1.rs` with substitutions:
+- `const VERSION_LABEL: &str = "code-go-ast-v1";`
+- struct name `CodeGoAstV1Chunker`
+- error message says `"CodeGoAstV1Chunker only handles..."`
+- module doc-comment prose `Rust` → `Go`, `code-rust-ast-v1` → `code-go-ast-v1`
+
+`split_oversize` / `make_chunk` / `AST_CHUNK_MAX_LINES = 200` / `BYTES_PER_TOKEN = 3` / `POLICY_HASH_HEX_LEN = 16` IDENTICAL (language-agnostic).
+
+Test module: copy from `code_ts_ast_v1.rs` and substitute names. KEEP cross-chunker `policy_hash_matches_md_heading_v1`.
+
+Wire into `crates/kebab-chunk/src/lib.rs`:
+
+```rust
+mod code_go_ast_v1;
+pub use code_go_ast_v1::CodeGoAstV1Chunker;
+```
+
+(Alphabetical placement.)
+
+Verify + commit:
+- `cargo test -p kebab-chunk code_go_ast` PASS (~6 tests)
+- `cargo test -p kebab-chunk` full per-crate green
+- `cargo clippy -p kebab-chunk --all-targets -- -D warnings` clean
+
+```bash
+git add crates/kebab-chunk/
+git commit -m "feat(p10-1c-go): code-go-ast-v1 chunker (1:1 + oversize split)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task F: Activate Go in app dispatch
+
+**Files:**
+- Modify: `crates/kebab-app/src/lib.rs` (replace 4 "go" bail! arms with real calls)
+- Modify: `crates/kebab-app/tests/code_ingest_smoke.rs` (add Go integration test)
+
+Replace the 4 `"go" => anyhow::bail!(...)` arms in `ingest_one_code_asset` (added in Task C) with real:
+
+```rust
+"go" => ParserVersion(kebab_parse_code::GO_PARSER_VERSION.to_string()),
+// ...
+"go" => CodeGoAstV1Chunker.chunker_version(),
+// ...
+"go" => kebab_parse_code::GoAstExtractor::new()
+    .extract(&ctx, &bytes)
+    .context("kb-parse-code::GoAstExtractor::extract (code:go)")?,
+// ...
+"go" => CodeGoAstV1Chunker
+    .chunk(&canonical, chunk_policy)
+    .context("kb-chunk::CodeGoAstV1Chunker::chunk (code:go)")?,
+```
+
+Add imports at top of lib.rs:
+- `kebab_chunk::CodeGoAstV1Chunker`
+- `kebab_parse_code::GoAstExtractor`
+
+Integration test (mirror PR #142's `python_file_ingests_and_searches_as_code_citation`):
+
+```rust
+#[test]
+fn go_file_ingests_and_searches_as_code_citation() {
+    // ... TempDir + Config harness same as Python/TS test ...
+    let pkg_dir = env.workspace_root.join("chunk");
+    std::fs::create_dir_all(&pkg_dir).unwrap();
+    std::fs::write(
+        pkg_dir.join("ast.go"),
+        "package chunk\n\nfunc ParseDoc(input string) string {\n    return input\n}\n",
+    ).unwrap();
+
+    let report = kebab_app::ingest_with_config(/* ... */).unwrap();
+    assert!(report.new >= 1);
+    let go_item = report.items.as_ref().unwrap().iter()
+        .find(|i| i.doc_path.0.ends_with("ast.go")).expect("ast.go item");
+    assert_eq!(go_item.parser_version.as_ref().unwrap().0, "code-go-v1");
+    assert_eq!(go_item.chunker_version.as_ref().unwrap().0, "code-go-ast-v1");
+
+    let hits = kebab_app::search_with_config(/* search "ParseDoc" */).unwrap();
+    let h = hits.iter().find(|h| matches!(h.citation, kebab_core::Citation::Code { .. }))
+        .expect("Citation::Code hit");
+    match &h.citation {
+        kebab_core::Citation::Code { lang, symbol, line_start, .. } => {
+            assert_eq!(lang.as_deref(), Some("go"));
+            assert_eq!(symbol.as_deref(), Some("chunk.ParseDoc"));
+            assert!(*line_start >= 1);
+        }
+        _ => unreachable!(),
+    }
+    assert_eq!(h.code_lang.as_deref(), Some("go"));
+}
+```
+
+Verify:
+- `cargo test -p kebab-app --test code_ingest_smoke` → 7/7 (6 existing + 1 new go)
+- `cargo test -p kebab-app --lib` → 52/52 (no regression)
+- clippy clean
+
+```bash
+git add crates/kebab-app/
+git commit -m "feat(p10-1c-go): activate Go in ingest_one_code_asset dispatch
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task G: Snapshot + full-suite gate + manual SMOKE
+
+**Files:**
+- Create: `crates/kebab-chunk/tests/code_go_ast_snapshot.rs` + fixture + baseline (mirror `code_python_ast_snapshot.rs` from PR #142)
+
+- [ ] **Step 1**: Add snapshot integration test. In-memory `CanonicalDocument` (no kebab-parse-code dep — boundary §6.3). Generate baseline: `UPDATE_SNAPSHOTS=1 cargo test -p kebab-chunk code_go_ast_snapshot` → re-run without env → PASS.
+
+- [ ] **Step 2**: Full-suite gate (the ONE invocation allowed this PR):
+
+```bash
+cargo clippy --workspace --all-targets -- -D warnings
+cargo test --workspace --no-fail-fast -j 1
+```
+
+Both must be CLEAN/GREEN.
+
+- [ ] **Step 3**: Manual SMOKE (optional but recommended — mirror PR #142 SMOKE):
+
+```bash
+cargo build --release   # OR debug if RAM-tight
+rm -rf /tmp/kebab-go-smoke && mkdir -p /tmp/kebab-go-smoke/ws/chunk
+echo 'package chunk
+
+func ParseDoc(input string) string { return input }
+' > /tmp/kebab-go-smoke/ws/chunk/ast.go
+# adapt isolated config from docs/SMOKE.md
+./target/release/kebab --config /tmp/kebab-go-smoke/config.toml ingest --json | jq '.items[].parser_version' | sort -u
+./target/release/kebab --config /tmp/kebab-go-smoke/config.toml search "ParseDoc" --code-lang go --json | jq '.hits[0]'
+```
+
+Expected: `code-go-v1` in parser_versions; Citation::Code with symbol `chunk.ParseDoc`.
+
+- [ ] **Step 4**: Commit snapshot only (full-suite + SMOKE are gates, not commit content):
+
+```bash
+git add crates/kebab-chunk/tests/
+git commit -m "test(p10-1c-go): code-go-ast-v1 chunker snapshot + full-suite gate
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task H: Docs + version bump
+
+- README: 지원 형식 row — add Go (`.go`, `code-go-ast-v1`).
+- HANDOFF: P10 phase row note 1C-Go merged (Go active). Java/Kotlin remain pending.
+- ARCHITECTURE: directory tree note for kebab-parse-code includes `go.rs` (Java/Kotlin coming in next PR). Decisions table — no new row (1C-Go follows the 1A-2/1B convention).
+- SMOKE: extend the P10 section with a 1-line note for Go (or compact Go example).
+- tasks/INDEX + tasks/p10/INDEX: flip the row for 1C-Go to 🟡 (PR open) → ✅ on merge. The 1C row in p10/INDEX may need a split — `p10-1C-Go ⏳ → 🟡` and `p10-1C-JavaKotlin ⏳ unchanged` (since user split into 2 PRs).
+- frozen design §10.1: add a one-liner — "p10-1C-Go 활성화 (Go)" (Java/Kotlin will get its own line in the next PR).
+- `Cargo.toml`: workspace version `0.11.1 → 0.12.0` (minor — dogfooding surface 확장, 새 chunker + extractor 활성화).
+
+```bash
+git add -A
+git commit -m "docs(p10-1c-go): README/HANDOFF/ARCHITECTURE/SMOKE/INDEX + chore: bump version 0.11.1 → 0.12.0
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Finalize: PR + review loop + release
+
+Per workflow memory (gitea-pr + review loop, no single-shot):
+
+- [ ] `gitea-pr` → PR title `feat(p10-1C-Go): tree-sitter-go AST extractor + chunker — Go 코드 색인 활성화`
+- [ ] Review loop until APPROVE → merge → main pull → branch cleanup → `cargo clean` → `gitea-release v0.12.0`.
+
+---
+
+## Self-Review (filled by plan author)
+
+- **Spec coverage**: design §1C Go (extractor + chunker + activation) → Tasks D/E/F; §3.3 (`code-go-ast-v1`) → Task E; §3.4 symbol path → Task D (extract_package + method receiver pointer detection); §6.1 (`kebab-parse-code/src/go.rs`) → Task D; §6.2 (`kebab-chunk/src/code_go_ast_v1.rs`) → Task E; §6.3 dep graph (`tree-sitter-go` parser-side) → Task A; §9.1 Tier-1 + oversize fallback → Task E (1A-2 split_oversize reused identically).
+- **No placeholders**: novel logic (`extract_package`, method receiver pointer detection, fixture, test assertions, dispatch arm additions) given concretely. Mechanical mirrors (chunker, integration test, snapshot test) pinned to exact existing files with substitutions.
+- **Type consistency**: `GoAstExtractor` / `GO_PARSER_VERSION = "code-go-v1"` / `CodeGoAstV1Chunker` / `VERSION_LABEL = "code-go-ast-v1"` used consistently across Tasks A-H. `MediaType::Code("go")` in routing + dispatch. `Citation::Code` with `lang: Some("go")` in integration test.

docs/superpowers/plans/2026-05-20-p10-1c-jk-ast-chunker.md

@@ -0,0 +1,494 @@
+# p10-1C-JavaKotlin Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development to implement this plan task-by-task.
+
+**Goal:** Activate Java + Kotlin code ingest end-to-end. Mirror 1C-Go (PR #151 / v0.12.0) for Java (single-language scaffold) and Kotlin (additional top-level fn variant). Both use source-side `package` extraction (design §3.4 JVM convention).
+
+**Architecture:** Same shape as 1B (multi-language single PR). 2 new tree-sitter grammars + 2 extractors + 2 chunkers + media routing + app dispatch arms. 1C-Go pattern is the closest template for source-side `package` extraction.
+
+**Tech Stack:** Rust 2024 workspace, `tree-sitter` 0.26 (already), `tree-sitter-java` + `tree-sitter-kotlin` (NEW). 1A-2/1B/1C-Go infrastructure unchanged.
+
+**Memory note:** Host has been OOM'd previously. Per-crate cargo only. ONE full-suite + clippy invocation in Task J.
+
+---
+
+## Pre-flight
+
+Branch `feat/p10-1c-jk` already exists.
+
+- [ ] **Disk hygiene**: `cargo clean` if heavy (last cleanup recovered 34 GB).
+
+Reference files:
+- 1C-Go extractor: `crates/kebab-parse-code/src/go.rs` — closest template for source-side package extraction.
+- 1B Python extractor: `crates/kebab-parse-code/src/python.rs` — class-nesting recursion model (relevant for Java/Kotlin).
+- 1A-2 chunker: `crates/kebab-chunk/src/code_rust_ast_v1.rs` — duplicate-with-substitution.
+- 1B dispatch generalization: `crates/kebab-app/src/lib.rs::ingest_one_code_asset` 4-arm match (~L1645). 1C-Go already added `"go"`; this PR adds `"java"` + `"kotlin"`.
+
+---
+
+## Task A: Workspace deps (tree-sitter-java + tree-sitter-kotlin)
+
+**Files:**
+- Modify: `Cargo.toml` (workspace `[workspace.dependencies]`, after `tree-sitter-go` line)
+- Modify: `crates/kebab-parse-code/Cargo.toml`
+
+- [ ] **Step 1**: `cargo add tree-sitter-java tree-sitter-kotlin -p kebab-parse-code`. If `tree-sitter-kotlin` resolves to a fork name, verify the actively-maintained crate (e.g. check crates.io page / GitHub stars / last update). Likely `tree-sitter-kotlin` (without fork suffix) is the default.
+
+- [ ] **Step 2**: Lift the two resolved versions into `[workspace.dependencies]` after `tree-sitter-go`:
+
+```toml
+# JVM family grammars for code ingest (kebab-parse-code, p10-1C-JK).
+tree-sitter-java       = "<resolved>"
+tree-sitter-kotlin     = "<resolved>"
+```
+
+Switch crate's entries to `{ workspace = true }`.
+
+- [ ] **Step 3**: `cargo build -p kebab-parse-code` → clean. Unused dep warning is fine.
+
+- [ ] **Step 4**: Commit:
+
+```bash
+git add Cargo.toml Cargo.lock crates/kebab-parse-code/Cargo.toml
+git commit -m "build(p10-1c-jk): add tree-sitter-java + tree-sitter-kotlin workspace deps
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+If the kotlin crate has a different actual name (e.g. `tree-sitter-kotlin-ng` or fork suffix), document the choice in the commit body briefly.
+
+---
+
+## Task B: source-fs routing `.java` / `.kt` / `.kts`
+
+**Files:**
+- Modify: `crates/kebab-source-fs/src/media.rs` (add arm after the existing `.go` arm)
+- Test: same file's test module
+
+- [ ] **Step 1 (failing test)** — add near `go_files_map_to_media_code_go`:
+
+```rust
+#[test]
+fn java_kotlin_files_map_to_media_code() {
+    assert_eq!(media_type_for(Path::new("a/b.java")), MediaType::Code("java".into()));
+    assert_eq!(media_type_for(Path::new("a/b.kt")), MediaType::Code("kotlin".into()));
+    assert_eq!(media_type_for(Path::new("a/b.kts")), MediaType::Code("kotlin".into()));
+}
+```
+
+- [ ] **Step 2**: Run → FAIL.
+
+- [ ] **Step 3**: Add the arms before the `_ => MediaType::Other(ext)` fallback (after `"go" => ...`):
+
+```rust
+        // p10-1C-JK: JVM family (Java + Kotlin) ingest activated.
+        "java"             => MediaType::Code("java".into()),
+        "kt" | "kts"       => MediaType::Code("kotlin".into()),
+```
+
+- [ ] **Step 4**: Run → PASS. `cargo test -p kebab-source-fs` → no regression.
+
+- [ ] **Step 5**: clippy clean, commit.
+
+```bash
+cargo clippy -p kebab-source-fs --all-targets -- -D warnings
+git add crates/kebab-source-fs/
+git commit -m "feat(p10-1c-jk): route .java/.kt/.kts to MediaType::Code
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task C: App dispatch + bail arms for "java" + "kotlin"
+
+**Files:**
+- Modify: `crates/kebab-app/src/lib.rs`
+
+- [ ] **Step 1**: Find the dispatch arm guard (currently `matches!(lang.as_str(), "rust" | "python" | "typescript" | "javascript" | "go")`). Add `"java"` + `"kotlin"`:
+
+```rust
+MediaType::Code(lang)
+    if matches!(lang.as_str(),
+        "rust" | "python" | "typescript" | "javascript" | "go" | "java" | "kotlin") =>
+```
+
+- [ ] **Step 2**: In `ingest_one_code_asset` the 4 `match code_lang` blocks add `"java"` and `"kotlin"` arms that `bail!()` for now:
+
+```rust
+"java" => anyhow::bail!("java ingest not yet wired (p10-1c-jk Task F)"),
+"kotlin" => anyhow::bail!("kotlin ingest not yet wired (p10-1c-jk Task I)"),
+```
+
+(in each of the 4 blocks before the `other =>` catch-all).
+
+- [ ] **Step 3**: Verify per-crate:
+- `cargo test -p kebab-app --lib` → 52 stay green
+- `cargo test -p kebab-app --test code_ingest_smoke` → 7 stay green
+- `cargo clippy -p kebab-app --all-targets -- -D warnings` clean
+
+- [ ] **Step 4**: Commit:
+
+```bash
+git add crates/kebab-app/
+git commit -m "refactor(p10-1c-jk): add java + kotlin to ingest dispatch allowlist (bail until Tasks F/I)
+
+Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Task D: `JavaAstExtractor`
+
+**Files:**
+- Create: `crates/kebab-parse-code/src/java.rs`
+- Modify: `crates/kebab-parse-code/src/lib.rs` (`pub mod java;` + re-exports `JAVA_PARSER_VERSION`, `JavaAstExtractor`)
+- Create: `crates/kebab-parse-code/tests/fixtures/sample.java`
+
+Scaffold mirrors `crates/kebab-parse-code/src/go.rs` (1C-Go) — single-language with source-side `package` extraction. Differences:
+
+### Constants
+
+```rust
+pub const PARSER_VERSION: &str = "code-java-v1";
+pub struct JavaAstExtractor;
+// supports: matches!(m, MediaType::Code(l) if l == "java")
+// code_lang = Some("java"), SourceType::Note, repo via detect_repo
+```
+
+### Package extraction (Java)
+
+tree-sitter-java grammar:
+- Root: `program`
+- `package_declaration` (top-level child) → contains `scoped_identifier` (dotted) OR `identifier` (single-segment)
+
+```rust
+fn extract_package(root: tree_sitter::Node, src: &str) -> Option<String> {
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        if child.kind() == "package_declaration" {
+            // package_declaration has scoped_identifier OR identifier as first named child
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                if sub.kind() == "scoped_identifier" || sub.kind() == "identifier" {
+                    return Some(src[sub.start_byte()..sub.end_byte()].to_string());
+                }
+            }
+        }
+    }
+    None
+}
+```
+
+(Verify field names against tree-sitter-java's node-types.json if any field differs.)
+
+### AST mapping
+
+| node kind | unit | symbol |
+|-----------|------|--------|
+| `class_declaration` (name field) | 1 + recurse body | `<pkg>.<ClassName>` |
+| `interface_declaration` (name) | 1 + recurse body | `<pkg>.<InterfaceName>` |
+| `enum_declaration` (name) | 1 | `<pkg>.<EnumName>` |
+| `record_declaration` (name, Java 14+) | 1 | `<pkg>.<RecordName>` |
+| `annotation_type_declaration` (name) | 1 | `<pkg>.<AnnotationName>` |
+| Inside class body: `method_declaration` (name) | 1 | `<pkg>.<Class>.<method>` |
+| Inside class body: `constructor_declaration` (name = class name) | 1 | `<pkg>.<Class>.<ClassName>` (matches Java convention) |
+| Nested classes recurse with class name pushed onto mod_path | as above | `<pkg>.<Outer>.<Inner>` etc. |
+| `import_declaration`, `package_declaration` | glue | `<pkg>.<top-level>` |
+| `field_declaration` at top of class | NOT a unit in 1C-JK (would explode unit count for value-only fields) | n/a |
+
+`unit_start` walks `comment` siblings; Java has `@interface` annotations but those are part of `annotation_type_declaration` itself, not separate sibling nodes.
+
+`mod_path` = class nesting (like 1B Python). Empty at file top level.
+
+### Fixture `tests/fixtures/sample.java`:
+
+```java
+// sample.java
+package com.kebab.chunk;
+
+import java.util.List;
+import java.util.stream.Collectors;
+
+/**
+ * Heading-aware Markdown chunker.
+ */
+public class MdHeadingV1Chunker {
+    private final String name;
+
+    public MdHeadingV1Chunker(String name) {
+        this.name = name;
+    }
+
+    public List<String> chunkDoc(String input) {
+        return List.of(name, input);
+    }
+
+    public String getName() {
+        return name;
+    }
+
+    public static class Builder {
+        private String name;
+        public Builder withName(String n) { this.name = n; return this; }
+        public MdHeadingV1Chunker build() { return new MdHeadingV1Chunker(name); }
+    }
+}
+
+interface Stringer {
+    String asString();
+}
+
+enum Mode { DEFAULT, FAST }
+```
+
+### Test module (inline `#[cfg(test)] mod tests`)
+
+Mirror 1C-Go shape:
+
+```rust
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use kebab_core::{Block, MediaType, SourceSpan};
+
+    fn extract_fixture() -> kebab_core::CanonicalDocument {
+        let bytes = std::fs::read(
+            concat!(env!("CARGO_MANIFEST_DIR"), "/tests/fixtures/sample.java"),
+        ).unwrap();
+        let asset = crate::rust::tests_support::fixed_code_asset(
+            "crates/x/src/sample.java", "java",
+        );
+        let cfg = kebab_core::ExtractConfig::default();
+        let root = std::path::PathBuf::from("/tmp");
+        let ctx = kebab_core::ExtractContext { asset: &asset, workspace_root: &root, config: &cfg };
+        JavaAstExtractor::new().extract(&ctx, &bytes).unwrap()
+    }
+
+    #[test]
+    fn extractor_supports_only_media_code_java() { /* ... */ }
+
+    #[test]
+    fn java_units_match_design_3_4_symbols() {
+        let doc = extract_fixture();
+        let mut syms: Vec<String> = doc.blocks.iter().filter_map(|b| match b {
+            Block::Code(c) => match &c.common.source_span {
+                SourceSpan::Code { symbol, lang, .. } => {
+                    assert_eq!(lang.as_deref(), Some("java"));
+                    symbol.clone()
+                }
+                _ => None,
+            },
+            _ => None,
+        }).collect();
+        syms.sort();
+        // workspace path → package extracted from source = com.kebab.chunk
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker"), "got {syms:?}");
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.MdHeadingV1Chunker"));  // constructor
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.chunkDoc"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.getName"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Builder"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Builder.withName"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.MdHeadingV1Chunker.Builder.build"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.Stringer"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.Mode"));
+        assert!(syms.iter().any(|s| s == "com.kebab.chunk.<top-level>"));
+    }
+
+    #[test]
+    fn deterministic_across_runs() {
+        let a = extract_fixture();
+        for _ in 0..50 { assert_eq!(extract_fixture().blocks, a.blocks); }
+    }
+}
+```
+
+### Wire into lib.rs
+
+```rust
+pub mod java;
+pub use java::{PARSER_VERSION as JAVA_PARSER_VERSION, JavaAstExtractor};
+```
+
+### Verify + commit
+
+- `cargo test -p kebab-parse-code` → all pass
+- `cargo clippy -p kebab-parse-code --all-targets -- -D warnings` clean
+- commit `feat(p10-1c-jk): tree-sitter-java AST extractor (JavaAstExtractor)`
+
+---
+
+## Task E: `code-java-ast-v1` chunker
+
+Identical pattern to 1C-Go Task E. Duplicate `code_rust_ast_v1.rs` with substitutions:
+- `VERSION_LABEL = "code-java-ast-v1"`, struct `CodeJavaAstV1Chunker`
+- error message + module doc-comment prose
+- Test module: parser_version `"code-java-v1"`, code_lang `"java"`
+- Keep cross-chunker `policy_hash_matches_md_heading_v1`
+
+Wire into `crates/kebab-chunk/src/lib.rs` (alphabetical). Verify + commit.
+
+---
+
+## Task F: Activate Java in app dispatch
+
+Replace the `"java"` `bail!()` arms in `ingest_one_code_asset` with real calls (`JavaAstExtractor` + `CodeJavaAstV1Chunker`). Add integration test `java_file_ingests_and_searches_as_code_citation` (mirror 1C-Go test, fixture `pkg_dir/Foo.java` with `package com.foo;` and `public class Foo { public String bar() { ... } }`, assert symbol `com.foo.Foo.bar`).
+
+Verify + commit.
+
+---
+
+## Task G: `KotlinAstExtractor`
+
+**Files:**
+- Create: `crates/kebab-parse-code/src/kotlin.rs`
+- Modify: `crates/kebab-parse-code/src/lib.rs`
+- Create: `crates/kebab-parse-code/tests/fixtures/sample.kt`
+
+Constants: `PARSER_VERSION = "code-kotlin-v1"`, `KotlinAstExtractor`, `code_lang = "kotlin"`.
+
+### Package extraction (Kotlin)
+
+tree-sitter-kotlin grammar:
+- Root: `source_file`
+- `package_header` (top-level) → contains `identifier` (dotted is single `identifier` node text; verify against node-types.json)
+
+```rust
+fn extract_package(root: tree_sitter::Node, src: &str) -> Option<String> {
+    let mut cur = root.walk();
+    for child in root.named_children(&mut cur) {
+        if child.kind() == "package_header" {
+            let mut c2 = child.walk();
+            for sub in child.named_children(&mut c2) {
+                if sub.kind() == "identifier" {
+                    return Some(src[sub.start_byte()..sub.end_byte()].to_string());
+                }
+            }
+        }
+    }
+    None
+}
+```
+
+(Verify against tree-sitter-kotlin's node-types.json — Kotlin grammar varies more than Java's.)
+
+### AST mapping (Kotlin)
+
+| node kind | unit | symbol |
+|-----------|------|--------|
+| `class_declaration` (name field) — covers `class`, `data class`, `sealed class`, `enum class`, `interface` (Kotlin's interface is a class_declaration variant) | 1 + recurse body | `<pkg>.<ClassName>` |
+| `object_declaration` (name) — singleton | 1 + recurse | `<pkg>.<ObjectName>` |
+| `function_declaration` (name) | 1 | `<pkg>.<fn_name>` (top-level) or `<pkg>.<Class>.<method>` (inside class) |
+| Inside class body: `function_declaration` → method | 1 | `<pkg>.<Class>.<method>` |
+| `property_declaration` at top-level (`val` / `var`) | glue | `<top-level>` (Kotlin top-level properties are common — keep as glue not unit) |
+| `import_header`, `package_header` | glue | `<top-level>` |
+
+(Detect class-vs-interface via modifier; for 1C 1차 treat both as `class_declaration` arm — symbol differs only via name. If tree-sitter-kotlin exposes `interface` keyword via modifier list, mention in HOTFIXES if special handling needed.)
+
+### Fixture `sample.kt`:
+
+```kotlin
+// sample.kt
+package com.kebab.chunk
+
+import java.util.List
+
+/**
+ * Heading-aware Markdown chunker.
+ */
+class MdHeadingV1Chunker(val name: String) {
+    fun chunkDoc(input: String): List<String> = listOf(name, input)
+
+    fun getName(): String = name
+
+    companion object {
+        fun withName(n: String): MdHeadingV1Chunker = MdHeadingV1Chunker(n)
+    }
+}
+
+interface Stringer {
+    fun asString(): String
+}
+
+enum class Mode { DEFAULT, FAST }
+
+fun freeFunction(x: Int): Int = x + 1
+
+object Singleton {
+    fun ping(): String = "pong"
+}
+```
+
+### Test module — assert symbols
+
+```rust
+// Asserted symbols:
+"com.kebab.chunk.MdHeadingV1Chunker"
+"com.kebab.chunk.MdHeadingV1Chunker.chunkDoc"
+"com.kebab.chunk.MdHeadingV1Chunker.getName"
+"com.kebab.chunk.MdHeadingV1Chunker.Companion"  // companion object (verify name)
+"com.kebab.chunk.MdHeadingV1Chunker.Companion.withName"  // method on companion
+"com.kebab.chunk.Stringer"
+"com.kebab.chunk.Mode"
+"com.kebab.chunk.freeFunction"  // top-level fn (Kotlin-specific!)
+"com.kebab.chunk.Singleton"
+"com.kebab.chunk.Singleton.ping"
+"com.kebab.chunk.<top-level>"  // import + property glue
+```
+
+(Companion object: tree-sitter-kotlin may use `companion_object` or `object_declaration` with `companion` modifier — verify and adjust the symbol if `Companion` isn't the right name.)
+
+### Wire into lib.rs
+
+```rust
+pub mod kotlin;
+pub use kotlin::{PARSER_VERSION as KOTLIN_PARSER_VERSION, KotlinAstExtractor};
+```
+
+Verify + commit.
+
+---
+
+## Task H: `code-kotlin-ast-v1` chunker
+
+Same pattern as Task E. Substitute kotlin labels. Verify + commit.
+
+---
+
+## Task I: Activate Kotlin in app dispatch
+
+Replace `"kotlin"` bail arms with real calls. Add integration test `kotlin_file_ingests_and_searches_as_code_citation`. Verify + commit.
+
+---
+
+## Task J: Snapshots + full-suite + SMOKE
+
+- Create 2 snapshot tests (`code_java_ast_snapshot.rs`, `code_kotlin_ast_snapshot.rs`) + baselines. Mirror 1C-Go Task G snapshot test.
+- ONE workspace test + clippy invocation.
+- Manual SMOKE: write a `.java` and `.kt` file in TempDir, ingest, search.
+
+Verify + commit (snapshot only).
+
+---
+
+## Task K: Docs + version bump
+
+- README + HANDOFF + ARCHITECTURE + SMOKE + 2 INDEX updates + design §10.1.
+- `Cargo.toml` version `0.12.0 → 0.13.0` (minor, surface 확장).
+
+Commit `docs(p10-1c-jk): ... + chore: bump 0.12.0 → 0.13.0`.
+
+---
+
+## Finalize
+
+`gitea-pr` → review loop → merge → main pull → branch cleanup → `cargo clean` → `gitea-release v0.13.0`.
+
+---
+
+## Self-Review (filled by plan author)
+
+- **Spec coverage**: design §1C Java + Kotlin → Tasks D-I; §3.4 symbol path → extractor (Java D, Kotlin G); §6.1/§6.2 module structure → Tasks D/E/G/H; §6.3 dep graph → Task A; §9.1 Tier-1 + oversize fallback → chunkers E/H.
+- **No placeholders**: novel logic (Java `extract_package`, Kotlin `extract_package`, AST walk arm tables) given concretely. Chunkers (E, H) are explicit "duplicate code_rust_ast_v1.rs with substitution X/Y/Z".
+- **Type consistency**: `JavaAstExtractor` / `JAVA_PARSER_VERSION` / `CodeJavaAstV1Chunker` + `KotlinAstExtractor` / `KOTLIN_PARSER_VERSION` / `CodeKotlinAstV1Chunker` used consistently. `MediaType::Code("java")` / `("kotlin")` in routing + dispatch.
+- **Kotlin grammar risk**: noted — tree-sitter-kotlin's exact node kinds (`class_declaration` vs `object_declaration`, `companion_object` vs companion modifier, `package_header` vs `package_directive`) should be verified against the resolved crate's node-types.json. Pin contract via test fixture; HOTFIXES any deviation found during implementation.

docs/superpowers/specs/2026-04-27-kebab-final-form-design.md

@@ -1543,6 +1543,12 @@ transitional 형태) 의 source of truth.
 
 **p10-1A-2 surface 활성화 (2026-05-19)**: Rust 소스코드 ingest (`code-rust-ast-v1` chunker, `tree-sitter-rust`) 가 활성화됨. `.rs` 파일을 워크스페이스에 두면 `kebab ingest` 가 AST 단위로 chunk 생성 + `citation.kind = "code"` 로 검색 가능. `kebab schema --json` 의 `stats.code_lang_breakdown` 에 `"rust": N` 이 표시됨. 본 activation 으로 kebab 자기 crate 를 dogfooding KB 에 색인 가능. `SourceSpan::Code` (§3.4) 와 `MediaType::Code` (§3.5) 는 1A-1 에서 이미 spec 에 반영됨. 두 deferred deviation (`AST_CHUNK_MAX_LINES` 상수 고정, `SourceType::Code` 미존재) 은 `tasks/HOTFIXES.md` (2026-05-19) 에 기록.
 
+**p10-1B 활성화 (Python / TypeScript / JavaScript) (2026-05-20)**: Python (`code-python-ast-v1`, `.py`), TypeScript (`code-ts-ast-v1`, `.ts`/`.tsx`), JavaScript (`code-js-ast-v1`, `.js`/`.mjs`/`.cjs`/`.jsx`) AST chunker 활성화. symbol path 는 workspace 경로 → module path prefix: Python = dotted (예: `kebab_eval.metrics.compute_mrr`), TypeScript/JavaScript = slash-style (예: `src/Foo.Foo.search`). Rust 1A-2 의 file-scope-only symbol 과 비일관 수용 (HOTFIXES 2026-05-20). expression-level 함수 (`const foo = () => {}`) 는 glue 처리 (HOTFIXES 2026-05-20).
+
+**p10-1C-Go 활성화 (Go) (2026-05-20)**: Go (`code-go-ast-v1`, `.go`) AST chunker 활성화. symbol = `<package>.<Func>` / `<package>.(*Receiver).<Method>` 형식.
+
+**p10-1C-JavaKotlin 활성화 (Java + Kotlin) (2026-05-20)**: Java (`code-java-ast-v1`, `.java`) + Kotlin (`code-kotlin-ast-v1`, `.kt`/`.kts`) AST chunker 활성화. symbol = `com.foo.Foo.bar` 형식 (패키지 + 클래스 + 메서드/필드). Kotlin grammar 은 `tree-sitter-kotlin-ng` 사용 (bare `tree-sitter-kotlin` 은 tree-sitter 0.21–0.23 고착으로 사용 불가).
+
 ### 10.2 MCP server transport (fb-30)
 
 `kebab mcp` 가 stdio JSON-RPC server. Rust SDK = `rmcp 1.6`. Tool surface

docs/wire-schema/v1/reset_report.schema.json

@@ -14,12 +14,18 @@
     "schema_version":           { "const": "reset_report.v1" },
     "scope":                    {
       "type": "string",
-      "enum": ["all", "data_only", "vector_only", "config_only"]
+      "enum": ["all", "data_only", "vector_only", "config_only", "orphans_only"]
     },
     "removed_paths":            {
       "type": "array",
       "items": { "type": "string" }
     },
-    "embedding_rows_truncated": { "type": "integer", "minimum": 0 }
+    "embedding_rows_truncated": { "type": "integer", "minimum": 0 },
+    "orphans_purged":           { "type": "integer", "minimum": 0, "default": 0 },
+    "purged_paths":             {
+      "type": "array",
+      "items": { "type": "string" },
+      "default": []
+    }
   }
 }

tasks/HOTFIXES.md

@@ -14,6 +14,40 @@ historical contract that was implemented; this file accumulates the
 deltas so phase 5+ readers can find the live behavior without diffing
 git history.
 
+## 2026-05-20 — p10-1B: Rust 1A-2 symbol path is file-scope-only; 1B+ uses workspace path → module prefix
+
+**무엇이 바뀌었나**: P10-1A-2 의 Rust `code-rust-ast-v1` chunker 가 생성하는 symbol 은 file-scope mod-path nesting 만 사용한다 (예: `Foo::double`). P10-1B 이후 Python / TypeScript / JavaScript 의 symbol 은 workspace 경로 → module path prefix 를 포함한다 (예: `kebab_eval.metrics.compute_mrr`, `src/Foo.Foo.search`).
+
+**원인**: 1A-2 는 symbol path 컨벤션이 확정되기 전에 구현됐고, 1B spec 에서 workspace path → module prefix 를 명시적 결정으로 확정했다 (p10-1b-py-ts-js-ast-chunkers.md §동결된 설계 결정). 1A-2 retrofit = `chunker_version` bump + Rust corpus 전체 re-ingest 비용이 수반됨.
+
+**사용자 가시적 영향**: Rust 코드 검색 시 symbol 이 `<ClassName>::<method>` 형태 (workspace prefix 없음). Python/TypeScript/JavaScript 는 `<module.path>.<symbol>` / `<module/path>.<symbol>` 형태. 비일관이지만 각각은 일관되게 동작.
+
+**proper fix**: Rust AST chunker 에 `module_path_for_rust(workspace_path)` helper 추가 + `chunker_version = "code-rust-ast-v2"` bump → 사용자가 명시 요청할 때까지 보류.
+
+**cross-link**: `tasks/p10/p10-1b-py-ts-js-ast-chunkers.md` Risks / notes 섹션, design §3.4.
+
+## 2026-05-20 — p10-1B: module_path_for_python / _tsjs do not sanitize non-ASCII / 공백 / 특수문자 in workspace path
+
+**동작**: `module_path_for_python` 와 `module_path_for_tsjs` 가 workspace path 의 비-ASCII / 공백 / 따옴표 / 백슬래시 같은 특수문자를 그대로 prefix 에 통과시킨다. 예: `kebab eval/metrics.py` (공백 포함) → module prefix `kebab eval.metrics` — 라이브러리 코드는 동작하지만 symbol 텍스트에 공백이 들어간다.
+
+**이유**: 1B 1차 단순화. 대다수 코드 베이스가 ASCII identifier + `/` 구분자만 사용하므로 사용자 경험상 영향 미미.
+
+**해결**: 후속 phase 에서 path-sanitize 추가 검토. NFKC normalize 후 `[^A-Za-z0-9_.\-/]` → `_` 변환 식. 적용 시 chunker_version bump 트리거 (re-ingest cascade 필요).
+
+**cross-link**: `tasks/p10/p10-1b-py-ts-js-ast-chunkers.md` Risks / notes 섹션 line 55.
+
+## 2026-05-20 — p10-1B: expression-level functions (arrow fn, function expression assigned to const) NOT emitted as units in 1B 1차
+
+**무엇이 바뀌었나**: TypeScript / JavaScript 의 `const foo = () => {...}` 또는 `const bar = function() {...}` 같은 expression-level 함수 할당은 `code-ts-ast-v1` / `code-js-ast-v1` 에서 독립 unit 으로 방출되지 않는다. 해당 코드는 가장 가까운 surrounding declaration-level unit (또는 `<top-level>` glue) 에 흡수된다.
+
+**원인**: `function_declaration` / `class_declaration` / `method_definition` / `interface_declaration` 같은 declaration-level 노드만 unit 으로 선택. `lexical_declaration` (= `const / let / var`) 안의 function / arrow expression 은 별도 unwrap 없이 pass-through. 1B 1차 단순화.
+
+**사용자 가시적 영향**: expression-level 함수 이름으로 검색 시 함수 body 를 포함하는 glue chunk 가 반환되지만, symbol 이 함수 이름 자체를 가리키지는 않는다. 함수명이 함수 본문 텍스트에 등장하므로 lexical / hybrid 검색으로 일반적으로 찾을 수 있다.
+
+**proper fix**: `lexical_declaration` visitor 에서 binding value 가 `arrow_function` / `function` expression 인 경우 해당 identifier name 을 symbol 로 사용하는 unwrap 추가. 후속 phase 에서 검토.
+
+**cross-link**: `tasks/p10/p10-1b-py-ts-js-ast-chunkers.md` Risks / notes 섹션.
+
 ## 2026-05-19 — p10-1A-2: AST_CHUNK_MAX_LINES constant vs config deviation
 
 **무엇이 바뀌었나**: `kebab-chunk/src/code_rust_ast_v1.rs` 가 `IngestCodeCfg.ast_chunk_max_lines` config 값을 읽지 않고 모듈 상수 `AST_CHUNK_MAX_LINES = 200` 으로 고정함.

tasks/INDEX.md

@@ -140,9 +140,10 @@ P0~P5 는 직렬. P6~P9 는 P5 이후 병렬 가능.
 
 - P10 — [p10/](p10/) — code ingest (multi-task, sub-indexed in [p10/INDEX.md](p10/INDEX.md))
   - [p10-1A-1 code ingest framework](p10/p10-1a-1-code-ingest-framework.md) — ✅ 머지
-  - [p10-1A-2 Rust AST chunker](p10/p10-1a-2-rust-ast-chunker.md) — 🟡 PR 오픈 (코드 완성, 머지 대기)
-  - p10-1B Python + TS/JS AST chunkers — ⏳
-  - p10-1C Go + Java + Kotlin AST chunkers — ⏳
+  - [p10-1A-2 Rust AST chunker](p10/p10-1a-2-rust-ast-chunker.md) — ✅ 머지
+  - [p10-1B Python + TS/JS AST chunkers](p10/p10-1b-py-ts-js-ast-chunkers.md) — 🟡 PR 오픈 (코드 완성, 머지 대기)
+  - p10-1C-Go Go AST chunker — 🟡 PR 오픈 (v0.12.0, `code-go-ast-v1`)
+  - p10-1C-JavaKotlin Java + Kotlin AST chunkers — 🟢 PR 오픈 (v0.13.0, `code-java-ast-v1` / `code-kotlin-ast-v1`)
   - p10-1D C + C++ AST chunkers — ⏳
   - p10-2 Tier 2 resource-aware — ⏳
   - p10-3 Tier 3 paragraph + line-window fallback — ⏳

tasks/p10/INDEX.md

@@ -3,9 +3,10 @@
 | ID | Subject | Status |
 |----|---------|--------|
 | 1A-1 | code ingest framework (wire schema, parse-code crate skeleton, filter flags, skip policy, config 절) | ✅ 머지 |
-| 1A-2 | Rust AST chunker | 🟡 PR 오픈 (코드 완성, 머지 대기) |
-| 1B | Python + TS/JS AST chunkers | ⏳ |
-| 1C | Go + Java + Kotlin AST chunkers | ⏳ |
+| 1A-2 | Rust AST chunker | ✅ 머지 |
+| 1B | Python + TS/JS AST chunkers | 🟡 PR 오픈 (코드 완성, 머지 대기) |
+| 1C-Go | Go AST chunker (`code-go-ast-v1`) | 🟡 PR 오픈 (v0.12.0) |
+| 1C-JavaKotlin | Java + Kotlin AST chunkers (`code-java-ast-v1` / `code-kotlin-ast-v1`) | 🟢 PR 오픈 (v0.13.0) |
 | 1D | C + C++ AST chunkers | ⏳ |
 | 2 | Tier 2 resource-aware (k8s / Dockerfile / manifest) | ⏳ |
 | 3 | Tier 3 paragraph + line-window fallback | ⏳ |

tasks/p10/p10-1b-py-ts-js-ast-chunkers.md

@@ -0,0 +1,60 @@
+# p10-1B — Python + TS/JS AST chunkers
+
+**Status:** 🟡 진행 중
+**Contract sections:** §3.3 (chunker_version `code-python-ast-v1` / `code-ts-ast-v1` / `code-js-ast-v1`), §3.4 (symbol path — Python `pkg.module.Class.method`, TS/JS `module/Class.method` / `module/default`), §3.5 (code_lang `python` / `typescript` / `javascript`), §5 (확장자 라우팅 활성화), §6.1 (`kebab-parse-code/src/{python,typescript,javascript}.rs`), §6.2 (`kebab-chunk/src/code_{python,ts,js}_ast_v1.rs`), §9.1 (Tier 1 AST per-language + oversize fallback).
+**Design:** [2026-05-15-kebab-code-ingest-design.md](../../docs/superpowers/specs/2026-05-15-kebab-code-ingest-design.md) §1B.
+**Plan:** [2026-05-20-p10-1b-py-ts-js-ast-chunkers.md](../../docs/superpowers/plans/2026-05-20-p10-1b-py-ts-js-ast-chunkers.md).
+
+## Goal
+
+1A-2 가 깐 인프라 (`SourceSpan::Code`, `MediaType::Code(String)`, `Citation::Code` 매핑, `citation_helper` arm, `backfill_code_lang` + `backfill_repo`, `schema.v1.code_lang_breakdown`, `[ingest.code]` 절, HOTFIXES) 위에 **Python + TypeScript + JavaScript** 3 언어의 extractor + chunker 를 활성화. design §1B 묶음과 일치하는 단일 PR. 머지 시점부터 Python / TS / JS 프로젝트도 dogfooding 가능.
+
+## 동결된 설계 결정 (이 task 로 확정)
+
+- **Symbol path 의 module prefix = workspace 경로 → module path 변환** (design §3.4 예시 충실, 사용자 명시 결정):
+  - **Python**: `crates/x/src/foo/bar.py` 같은 workspace_path 를 `/`/`__init__.py` 처리 + `.py`·`.pyi` strip + `/` → `.` 변환 후 dotted prefix 로 사용. 예시: `kebab_eval/metrics.py` 의 `def compute_mrr()` → symbol `kebab_eval.metrics.compute_mrr`. `pkg/__init__.py` 는 module `pkg` 자체. 변환은 `kebab-parse-code::lang::module_path_for_python(workspace_path)` 단일 함수 (source of truth).
+  - **TS/JS**: `src/search/retriever/Retriever.ts` → `src/search/retriever/Retriever` prefix + `/` 구분자 보존 + `.ts`/`.tsx`/`.js`/`.jsx`/`.mjs`/`.cjs` strip. 예시: `src/search/retriever/Retriever.ts` 의 method `search` → `src/search/retriever/Retriever.search`. `module/default` 는 `export default function/class` 경우. 변환은 `module_path_for_tsjs(workspace_path)`.
+  - **Rust 1A-2 는 retrofit 하지 않음** — 1A 는 file-scope nesting 만 사용 (workspace prefix 없음). 비일관 수용; HOTFIXES 2026-05-20 에 기록 + 사용자가 명시 요청 시 retrofit (chunker_version bump + re-ingest cascade 필요).
+- **TypeScript grammar selection**: `tree-sitter-typescript` crate 의 `LANGUAGE_TYPESCRIPT` 는 `.ts`, `LANGUAGE_TSX` 는 `.tsx` 에 사용. 파일 확장자로 선택. `code-ts-ast-v1` 하나의 chunker 가 둘 다 처리 (parser_version `code-ts-v1`).
+- **JavaScript grammar**: `tree-sitter-javascript` 단일 LanguageFn 가 `.js` / `.mjs` / `.cjs` / `.jsx` 모두 처리. 별도 분기 불필요.
+- **Expression-level 함수 (arrow fn / function expression assigned to const)**: 1B 1차에서는 *declaration-level 만* unit (function_declaration / class_declaration / method_definition / interface_declaration / type_alias_declaration / decorated_definition 등). `const foo = () => {...}` 같은 expression-level 은 glue 로 잡힘. HOTFIXES 2026-05-20 기록; 후속 phase 에서 lexical_declaration 안의 함수 표현식 unwrap 추가 검토.
+- **App dispatch 일반화**: 현재 `ingest_one_code_asset` 은 RustAstExtractor + CodeRustAstV1Chunker 하드코딩. 1B 에서 `lang: &str` 받아 dispatch (Rust 도 동일 함수로 흡수) — Extractor 와 Chunker 를 trait object 가 아니라 enum/match 로 선택 (kebab-app 만 변경, kebab-core/Chunker trait 불변). frozen design 영향 없음.
+- frozen design 자체는 변경 없음 (§3.4 의 symbol path 예시는 이미 본 결정과 일치). §10.1 (post-merge surface) 에 1B 활성화 한 줄 추가.
+
+## Acceptance criteria
+
+- `cargo test --workspace --no-fail-fast -j 1` passes (메모리 의식적으로는 per-crate; full-suite gate 는 Task K 직전 1회).
+- `cargo clippy --workspace --all-targets -- -D warnings` passes.
+- 3 언어 각각의 fixture (`tests/fixtures/sample.{py,ts,js}`) ingest → chunk snapshot 안정 + `Citation::Code` 의 symbol/line 이 §3.4 컨벤션 (workspace path → module path) 과 일치.
+- 격리 TempDir KB 에 Python/TS/JS 파일 하나씩 두고 `kebab search --code-lang {python|typescript|javascript} --json` 가 정상 결과 반환.
+- `kebab schema --json | jq .stats.code_lang_breakdown` 에 `python`, `typescript`, `javascript` 카운트 등장.
+- README + HANDOFF + ARCHITECTURE + SMOKE + tasks/INDEX + tasks/p10/INDEX 갱신.
+- frozen design §10.1 한 줄 추가 (1B 활성화).
+- HOTFIXES 2026-05-20 에 (a) Rust 1A-2 symbol path 비일관 (1B 와 다름), (b) expression-level 함수 단위 제외 — 두 편차 기록.
+- workspace `Cargo.toml` minor bump (0.7.0 → 0.8.0) — 도그푸딩 가능 surface 확장.
+
+## Allowed dependencies
+
+- `kebab-parse-code` 에 `tree-sitter-python`, `tree-sitter-typescript`, `tree-sitter-javascript` 추가 (workspace deps 경유). 기존 `kebab-core` / `anyhow` / `gix` / `tree-sitter` / `tree-sitter-rust` / `serde_json` / `time` / `tracing` 유지.
+- `kebab-chunk` 의 새 모듈 3개 (`code_python_ast_v1.rs` / `code_ts_ast_v1.rs` / `code_js_ast_v1.rs`) — 1A-2 chunker 와 동일 dep (kebab-core + serde_json_canonicalizer + blake3 + anyhow + tracing). tree-sitter 절대 import 금지.
+- `kebab-app` 변경 — 새 crate dep 없음.
+- `kebab-source-fs` — 확장자 추가만, 새 dep 없음.
+
+## Forbidden dependencies
+
+- `kebab-chunk` 가 `tree-sitter-*` 직접 import 금지 (AST 는 parser-side).
+- UI crate (cli / mcp / tui) 가 `kebab-parse-code` 직접 import 금지.
+- `kebab-parse-code` 가 store / embed / llm / rag 직접 import 금지 (design §8 inheritance).
+
+## Risks / notes
+
+- tree-sitter-typescript 의 `LANGUAGE_TYPESCRIPT` 와 `LANGUAGE_TSX` 가 별도 LanguageFn — 잘못 선택하면 TSX JSX 가 parse 실패. 파일 확장자 기반 선택을 단일 함수에서 결정 (테스트로 고정).
+- tree-sitter-python 의 `decorated_definition` 노드 처리 — 데코레이터가 wrap 하는 형태라 `function_definition` / `class_definition` 가 child. unwrap 필요 (decorator 라인은 unit_start backward extension 으로 자연스럽게 포함됨).
+- Python `pkg/__init__.py` 의 module path = `pkg` 자체 (basename 제거). `module_path_for_python` 가 이걸 처리.
+- TS/JS 의 `export default function/class` — name 이 없을 수 있음 (`export default function () {...}`). symbol `module/default` 로 표기 (design §3.4).
+- `module_path_for_python` / `module_path_for_tsjs` 가 workspace_path 의 비-ASCII / 공백 / 특수문자 처리 필요. 1B 1차에서는 그대로 전달 (sanitize 없음); HOTFIXES 에 path-sanitize 부재 기록.
+- 1A-2 `ingest_one_code_asset` 일반화로 인한 dispatch 코드 변경 — Rust 기존 동작 byte-identical 유지를 통합 테스트로 확인.
+- 머지 후 deviation 은 `tasks/HOTFIXES.md` 에 dated 로그 + 본 spec `Risks / notes` 에 one-line cross-link.
+- **[HOTFIXES 2026-05-20]** Rust 1A-2 symbol 은 file-scope nesting 만 (workspace prefix 없음); 1B 의 Python/TypeScript/JavaScript 와 비일관 — retrofit 은 사용자 명시 요청 시. 자세한 내용: `tasks/HOTFIXES.md` (2026-05-20, "Rust 1A-2 symbol path").
+- **[HOTFIXES 2026-05-20]** TypeScript/JavaScript 의 expression-level 함수 (`const foo = () => {}` 등) 는 `<top-level>` glue 로 처리됨, 독립 unit 미방출 — 후속 phase 에서 `lexical_declaration` unwrap 검토. 자세한 내용: `tasks/HOTFIXES.md` (2026-05-20, "expression-level functions").
+- **[HOTFIXES 2026-05-20]** `module_path_for_python` / `module_path_for_tsjs` 가 path-sanitize 안 함 (특수문자/공백 그대로 prefix 에 들어감) — 후속 phase 에서 NFKC + 사용금지 문자 변환 검토. 자세한 내용: `tasks/HOTFIXES.md` (2026-05-20, "module_path_for_python / _tsjs do not sanitize").

tasks/p10/p10-1c-go-ast-chunker.md

@@ -0,0 +1,54 @@
+# p10-1C-Go — Go AST chunker
+
+**Status:** 🟡 진행 중
+**Contract sections:** §3.3 (chunker_version `code-go-ast-v1`), §3.4 (symbol path — Go `package.Func` / `package.(*Receiver).Method`), §3.5 (code_lang `go`, ext `.go`), §6.1 (`kebab-parse-code/src/go.rs`), §6.2 (`kebab-chunk/src/code_go_ast_v1.rs`), §9.1 (Tier 1 AST per-language + oversize fallback).
+**Design:** [2026-05-15-kebab-code-ingest-design.md](../../docs/superpowers/specs/2026-05-15-kebab-code-ingest-design.md) §1C (Go 부분 — Java + Kotlin 은 후속 PR).
+**Plan:** [2026-05-20-p10-1c-go-ast-chunker.md](../../docs/superpowers/plans/2026-05-20-p10-1c-go-ast-chunker.md).
+
+## Goal
+
+1A-2 / 1B 인프라 위에 Go AST chunker 활성화. 사용자 결정으로 1C 의 3 언어 (Go + Java + Kotlin) 를 2 PR 로 분할 — Go 가 method receiver / package convention 면에서 Java/Kotlin (JVM family) 과 다르므로 별 PR. 본 PR 머지 시점부터 Go 프로젝트 dogfooding 가능.
+
+## 동결된 설계 결정 (이 task 로 확정)
+
+- **Symbol path 의 package prefix = 소스 코드의 `package` 선언에서 추출** (design §3.4 그대로). 1B 의 workspace-path 변환과 다름 — Go 는 언어 자체에 `package` declaration 이 있어 그게 canonical source. tree-sitter-go 의 `source_file` root 의 첫 named child `package_clause` 에서 추출. 빈 경우 (이론상 invalid Go, 실용엔 거의 없음) `<unknown>` 또는 fallback `<package>` (1A `<module>` 패턴과 유사).
+- **Method receiver 표현** (design 예시 그대로): `package.(*Receiver).Method` (포인터 receiver), `package.(Receiver).Method` (value receiver). tree-sitter-go 의 `method_declaration` 의 `receiver` field 에서 type + pointer 여부 추출. 예: `func (m *MdHeadingV1Chunker) ChunkDoc(...)` → symbol `chunk.(*MdHeadingV1Chunker).ChunkDoc`.
+- **Top-level unit 종류**:
+  - `function_declaration` → 1 unit, symbol `package.Func`
+  - `method_declaration` → 1 unit, symbol `package.(*Receiver).Method` / `package.(Receiver).Method`
+  - `type_declaration` (struct / interface / type alias) → 1 unit each, symbol `package.TypeName`
+  - `const_declaration`, `var_declaration`, `import_declaration` (블록 또는 단일) → glue, grouped → `package.<top-level>` (1A/1B 패턴)
+- **Go 의 generic 처리**: `func Foo[T any](...)` 또는 `type Foo[T any] struct{}` 의 type parameter 는 symbol 에 미포함 (Go 자체도 보통 symbol 에 안 적음). 단순 `package.Foo` 만.
+- **Test detection**: Go 의 `func TestXxx(t *testing.T)` 는 *일반 fn 으로 emit*. test 감지 boost/penalty 등 ranking 영향은 본 task 범위 밖 (ranking brainstorm 보류 메모리 따름).
+- frozen design 자체는 변경 없음 (§3.4 의 Go 행이 이미 본 결정과 일치). §10.1 에 1C-Go 활성화 한 줄 추가.
+
+## Acceptance criteria
+
+- `cargo test --workspace --no-fail-fast -j 1` passes (memory-conscious: per-crate 위주, full-suite gate 는 docs task 직전 1회).
+- `cargo clippy --workspace --all-targets -- -D warnings` passes.
+- Go fixture (`tests/fixtures/sample.go`) ingest → chunk snapshot 안정 + `Citation::Code` 의 symbol 이 §3.4 컨벤션 일치 (`pkg.Func` / `pkg.(*Receiver).Method`).
+- 격리 TempDir KB 에 Go 파일 두고 `kebab search --code-lang go --json` 가 `Citation::Code { lang: "go", symbol: "...", ... }` 반환.
+- `kebab schema --json | jq .stats.code_lang_breakdown` 에 `"go"` 카운트.
+- README + HANDOFF + ARCHITECTURE + SMOKE + tasks/INDEX + tasks/p10/INDEX 갱신.
+- frozen design §10.1 한 줄 추가.
+- workspace `Cargo.toml` minor bump (0.11.1 → 0.12.0).
+
+## Allowed dependencies
+
+- `kebab-parse-code` 에 `tree-sitter-go` 추가 (workspace deps). 기존 deps 유지.
+- `kebab-chunk` 의 새 모듈 `code_go_ast_v1.rs` — kebab-core + serde_json_canonicalizer + blake3 + anyhow + tracing. tree-sitter 절대 import 금지.
+- `kebab-app`, `kebab-source-fs` 변경 — 새 crate dep 없음.
+
+## Forbidden dependencies
+
+- `kebab-chunk` 가 `tree-sitter-go` 직접 import 금지.
+- UI crate 가 `kebab-parse-code` 직접 import 금지.
+- `kebab-parse-code` 가 store / embed / llm / rag 직접 import 금지.
+
+## Risks / notes
+
+- tree-sitter-go 의 `package_clause` node 가 root 의 첫 named child 인지 grammar 버전에 따라 다를 수 있음 — extractor 가 `source_file` 전체를 named_children iterate 하면서 첫 `package_clause` 잡는 방식이 안전.
+- `method_declaration` 의 receiver pointer 여부: tree-sitter-go AST 에서 receiver type 이 `pointer_type` 노드면 `*Receiver`, 그냥 `type_identifier` 면 `Receiver`. 정확한 텍스트 추출 필요.
+- Generic type parameter (`[T any]`) 가 method_declaration / function_declaration 의 name field 와 별도 child — name 만 추출하면 generic 부분 자동 제외.
+- 1B Python/TS/JS 패턴 (helpers from lang.rs) 와 *다른* 모델 — 본 task 의 mod_prefix 는 source-side AST 에서 추출, helper fn 불필요.
+- 머지 후 deviation 은 `tasks/HOTFIXES.md` 에 dated 로그 + 본 spec `Risks / notes` 에 one-line cross-link.

tasks/p10/p10-1c-jk-ast-chunker.md

@@ -0,0 +1,69 @@
+# p10-1C-JavaKotlin — Java + Kotlin AST chunkers
+
+**Status:** 🟡 진행 중
+**Contract sections:** §3.3 (chunker_version `code-java-ast-v1` + `code-kotlin-ast-v1`), §3.4 (symbol path — Java/Kotlin `package.Class.method`), §3.5 (code_lang `java` + `kotlin`, ext `.java` / `.kt` / `.kts`), §6.1 (`kebab-parse-code/src/{java,kotlin}.rs`), §6.2 (`kebab-chunk/src/code_{java,kotlin}_ast_v1.rs`), §9.1 (Tier 1 AST per-language + oversize fallback).
+**Design:** [2026-05-15-kebab-code-ingest-design.md](../../docs/superpowers/specs/2026-05-15-kebab-code-ingest-design.md) §1C (Java + Kotlin 부분 — Go 는 PR #151 / v0.12.0 별 PR 완료).
+**Plan:** [2026-05-20-p10-1c-jk-ast-chunker.md](../../docs/superpowers/plans/2026-05-20-p10-1c-jk-ast-chunker.md).
+
+## Goal
+
+1C-Go (PR #151 / v0.12.0) 의 자매 PR. 같은 1C phase 의 JVM family (Java + Kotlin) 묶음. 머지 시점부터 `.java` / `.kt` / `.kts` 파일 dogfooding 가능.
+
+## 동결된 설계 결정 (이 task 로 확정)
+
+- **Symbol prefix = 소스 코드의 `package` 선언에서 추출** (design §3.4 그대로, 1C-Go 모델과 동일). 1B 의 workspace-path 변환과 다름.
+  - **Java**: tree-sitter-java 의 `package_declaration` → 안의 `scoped_identifier` 또는 `identifier` 텍스트 (e.g. `com.kebab.chunk`). 없으면 `<unknown>`.
+  - **Kotlin**: tree-sitter-kotlin 의 `package_header` → `identifier` 텍스트. 없으면 (default package) `<unknown>`.
+- **Symbol 형식** (design §3.4): `package.Class.method`. 예시: `com.kebab.chunk.MdHeadingV1Chunker.chunkDoc`.
+- **Java AST mapping**:
+  - `class_declaration` (name) → 1 unit + recurse body
+  - `interface_declaration` (name) → 1 unit + recurse
+  - `enum_declaration` (name) → 1 unit
+  - `record_declaration` (Java 14+) (name) → 1 unit
+  - `annotation_type_declaration` → 1 unit
+  - Inside class/interface/enum: `method_declaration` (name) → unit `package.Class.method` (class nesting like 1B Python)
+  - `import_declaration`, `package_declaration` 자체 → glue `<top-level>` 
+  - Top-level fn 없음 (Java 자체에 없음)
+- **Kotlin AST mapping**:
+  - `class_declaration` (name) → 1 unit + recurse class_body. `data class` / `sealed class` / `enum class` 도 같은 노드.
+  - `object_declaration` (name) → 1 unit + recurse class_body (singleton)
+  - `function_declaration` (name) — **top-level 가능** → unit `package.fnName`. Class 내부면 `package.Class.method`.
+  - `property_declaration` at top-level → glue
+  - `interface` (in tree-sitter-kotlin 보통 `class_declaration` with `interface` modifier 또는 별 노드) → 1 unit
+  - `import_header`, `package_header` 자체 → glue `<top-level>`
+- **Glue grouping**: 1B Python / 1C-Go 패턴 동일 — imports + 기타 → 하나의 `<top-level>` (또는 `<module>` post-pass if file has zero real units).
+- **Tree-sitter Kotlin crate 선택**: tree-sitter-kotlin 의 가장 잘 유지되는 crate 사용 (`tree-sitter-kotlin` 또는 fork). resolve 시 active maintainer 확인.
+- frozen design 자체 변경 없음 — §10.1 에 1C-JK 활성화 한 줄.
+
+## Acceptance criteria
+
+- `cargo test --workspace --no-fail-fast -j 1` passes.
+- `cargo clippy --workspace --all-targets -- -D warnings` passes.
+- Java/Kotlin fixture 각각 (`tests/fixtures/sample.java`, `tests/fixtures/sample.kt`) ingest → chunk snapshot 안정 + symbol 이 §3.4 컨벤션 일치.
+- 격리 TempDir KB 에 `.java` / `.kt` 파일 두고 `kebab search --code-lang java --json` / `--code-lang kotlin --json` 가 `Citation::Code` 반환.
+- `kebab schema --json | jq .stats.code_lang_breakdown` 에 `"java"` + `"kotlin"` 카운트.
+- README + HANDOFF + ARCHITECTURE + SMOKE + tasks/INDEX + tasks/p10/INDEX 갱신.
+- frozen design §10.1 한 줄.
+- workspace `Cargo.toml` minor bump (0.12.0 → 0.13.0).
+
+## Allowed dependencies
+
+- `kebab-parse-code` 에 `tree-sitter-java` + `tree-sitter-kotlin` 추가. 기존 deps 유지.
+- `kebab-chunk` 의 새 모듈 2개 (`code_java_ast_v1.rs`, `code_kotlin_ast_v1.rs`) — language-agnostic body. tree-sitter import 금지.
+- `kebab-app`, `kebab-source-fs` — 새 crate dep 없음.
+
+## Forbidden dependencies
+
+- `kebab-chunk` 가 tree-sitter-java / tree-sitter-kotlin import 금지 (boundary §6.3).
+- UI crate 가 `kebab-parse-code` 직접 import 금지.
+- `kebab-parse-code` 가 store / embed / llm / rag 직접 import 금지.
+
+## Risks / notes
+
+- tree-sitter-kotlin: 공식 또는 가장 활발히 유지되는 crate (`tree-sitter-kotlin` 또는 fork) 선택 필요. resolve 시 metadata 확인.
+- Kotlin 의 grammar 가 다른 tree-sitter-* 보다 update 빈도 낮을 수 있어 grammar field 명 변동 가능 — 테스트 fixture 로 contract 고정.
+- Java record (Java 14+) — tree-sitter-java 에서 `record_declaration` 노드 (확인 필요).
+- Kotlin sealed class / data class / object declaration 등 변종 노드 — tree-sitter-kotlin 의 정확한 node kind 명 확인 필요 (grammar.json / node-types.json).
+- Java class 안의 inner class — Python 패턴 (recursion with class name pushed) 동일 처리.
+- Kotlin top-level fn 은 1B Python 의 top-level fn 패턴 + 1C-Go 의 package-prefix 패턴 hybrid — `package.fnName`.
+- 머지 후 deviation 은 `tasks/HOTFIXES.md` dated 로그 + 본 spec `Risks / notes` cross-link.