p1-3: add parse_blocks (pulldown-cmark walker) submodule

Implements `kb_parse_md::parse_blocks(body, body_offset_lines)` returning a
flat `Vec<ParsedBlock>` plus warnings. Walks pulldown-cmark events through a
small frame-based state machine that tracks heading paths, accumulates
inline buffers (Text/Code/Link/Strong/Emph only — design §3.4), and
reports SourceSpan::Line spans in 1-indexed file-line coordinates.

Covers headings, paragraphs, code blocks (lang from info string), GFM
tables (with malformed fallback to paragraph + MalformedTable warning),
lists (nested sub-lists flattened into parent item), and block-level image
references. Inline images are dropped silently per the inline filter.
Adversarial inputs are caught with `catch_unwind` and degrade to an empty
output + ExtractFailed warning.

15 unit tests cover heading-path correctness, code lang, table parsing,
malformed-table fallback (driven via synthetic events since pulldown-cmark
auto-normalizes table widths), LF/CRLF line-range parity, image refs,
nested-list flattening, inline filter, and 100-iteration random-bytes plus
hand-crafted adversarial-input no-panic guards.

Cargo.lock

@@ -585,11 +585,13 @@ dependencies = [
  "kb-core",
  "kb-parse-types",
  "lingua",
+ "pulldown-cmark",
  "serde",
  "serde_json",
  "serde_yaml_ng",
  "time",
  "toml",
+ "tracing",
 ]
 
 [[package]]
@@ -834,6 +836,17 @@ dependencies = [
  "unicode-ident",
 ]
 
+[[package]]
+name = "pulldown-cmark"
+version = "0.13.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7c3a14896dfa883796f1cb410461aef38810ea05f2b2c33c5aded3649095fdad"
+dependencies = [
+ "bitflags",
+ "memchr",
+ "unicase",
+]
+
 [[package]]
 name = "quote"
 version = "1.0.45"
@@ -1367,6 +1380,12 @@ dependencies = [
  "tracing-serde",
 ]
 
+[[package]]
+name = "unicase"
+version = "2.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dbc4bc3a9f746d862c45cb89d705aa10f187bb96c76001afab07a0d35ce60142"
+
 [[package]]
 name = "unicode-ident"
 version = "1.0.24"

crates/kb-parse-md/Cargo.toml

@@ -5,7 +5,7 @@ edition       = { workspace = true }
 rust-version  = { workspace = true }
 license       = { workspace = true }
 repository    = { workspace = true }
-description   = "Markdown frontmatter (and, in p1-3, block) parsing into kb-core::Metadata / kb-parse-types intermediates"
+description   = "Markdown frontmatter and block parsing into kb-core::Metadata / kb-parse-types intermediates"
 
 [dependencies]
 kb-core         = { path = "../kb-core" }
@@ -14,6 +14,12 @@ anyhow          = { workspace = true }
 serde           = { workspace = true }
 serde_json      = { workspace = true }
 time            = { workspace = true }
+tracing         = { workspace = true }
+# pulldown-cmark is the CommonMark parser used by the `blocks` submodule.
+# GFM tables are gated by the runtime `Options::ENABLE_TABLES` flag, not a
+# cargo feature; we strip the default `getopts` + `html` features since we
+# only use the pull-parser API.
+pulldown-cmark  = { version = "0.13", default-features = false }
 # serde_yaml (dtolnay) was archived as unmaintained in 2024.
 # We use the maintained fork serde_yaml_ng. Keeping the same `serde_yaml`-style
 # API surface lets us swap if a different fork wins long term.

crates/kb-parse-md/src/lib.rs

@@ -1,19 +1,21 @@
 //! `kb-parse-md` — Markdown parsing for the KB pipeline (§3.7b).
 //!
-//! P1-2 implements the **frontmatter** submodule only. P1-3 will add a
+//! Public surface:
-//! sibling `blocks` submodule for block parsing using `pulldown-cmark`.
-//!
-//! Public surface for P1-2 is intentionally narrow:
 //!
 //! * [`parse_frontmatter`] — pure function from Markdown bytes to
-//!   `(Metadata, Option<FrontmatterSpan>, Vec<Warning>)`.
+//!   `(Metadata, Option<FrontmatterSpan>, Vec<Warning>)` (P1-2).
 //! * [`BodyHints`] — caller-supplied fallbacks that feed the §0 Q9 derive
-//!   table when frontmatter is missing or partial.
+//!   table when frontmatter is missing or partial (P1-2).
 //! * [`FrontmatterSpan`] — byte offsets of the frontmatter region in the
-//!   input slice (returned by [`parse_frontmatter`]).
+//!   input slice (returned by [`parse_frontmatter`]) (P1-2).
+//! * [`parse_blocks`] — pure function from Markdown body bytes to
+//!   `(Vec<ParsedBlock>, Vec<Warning>)` with heading paths and 1-indexed
+//!   `SourceSpan::Line` ranges relative to the original file (P1-3).
 //!
 //! Anything else in this crate is `pub(crate)` and may change without notice.
 
+pub mod blocks;
 pub mod frontmatter;
 
+pub use blocks::parse_blocks;
 pub use frontmatter::{BodyHints, FrontmatterSpan, parse_frontmatter};