fix(ci): repair doc-truthing guard after #614's .md→.adoc migration (#617)

## What

Repairs `main`'s **red `CI`** after #614 merged. The `build` job's
`./tools/check-doc-truthing.sh` governance guard failed because **#614
migrated six status/history docs from `.md` → `.adoc`**, but the guard
still keyed off the old `.md` paths in two places:

- **Presence list** (`BANNERED_DOCS`) → `bannered doc is missing` for
`BACKEND-IMPLEMENTATION`, `COMPILER-CAPABILITIES`,
`ALPHA-1-RELEASE-NOTES`.
- **Over-claim ratchet baseline** (`tools/doc-overclaims.allow`) → `new
over-claim(s)` for the migrated `.adoc` history/roadmap snapshots
(identical, already-blessed claims — just at `.adoc` paths with `.adoc`
markup).

On the previous `main` tip (`2aa00ff`) this guard was green, so this is
a regression introduced purely by the migration — not new over-claiming.

## Fix (hotfix + hardening)

1. **Re-key the over-claim baseline** to the migrated `.adoc` paths via
the tool's documented `--update`. This is a **1:1 re-bless of identical
historical/roadmap content** — every removed `.md` signature has an
equivalent `.adoc` one added; **zero new live over-claims**.
2. **Harden the presence check** so it can't re-break on a future
migration: `BANNERED_DOCS` are now extension-less **stems** resolved to
whichever of `.adoc`/`.md` is present (`.adoc` preferred).

## Verification

- `./tools/check-doc-truthing.sh` → exit 0: *"OK: doc-truthing intact —
presence invariants + over-claim ratchet (DOC-04/05/08/09)."*
- Resolver unit-tested: both-present → `.adoc`; only `.md` → `.md`;
neither → missing.
- `dune build` clean. (The only `build`-job step after the guard is
`dune build @fmt`; this change touches no OCaml, so it's unaffected.)

Greens the `build` job / `CI` on `main`.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_01Lz7pRcec2Z3tVtaAhvB3M8

---
_Generated by [Claude
Code](https://claude.ai/code/session_01Lz7pRcec2Z3tVtaAhvB3M8)_

Co-authored-by: Claude <noreply@anthropic.com>

Author: hyperpolymath

tools/check-doc-truthing.sh

@@ -55,11 +55,14 @@ MATRIX="docs/CAPABILITY-MATRIX.adoc"
 
 # Docs that historically over-claim and MUST carry a pointer back at $MATRIX
 # (DOC-04). Each must mention CAPABILITY-MATRIX.adoc somewhere in its body.
+# Listed as extension-less stems and resolved to whichever of .adoc/.md is
+# present (.adoc preferred), so a .md<->.adoc migration (DOC-FORMAT) does not
+# break this presence check the way a hard-coded extension would.
 BANNERED_DOCS=(
-  "README.adoc"
-  "docs/architecture/BACKEND-IMPLEMENTATION.md"
-  "docs/reference/COMPILER-CAPABILITIES.md"
-  "docs/history/ALPHA-1-RELEASE-NOTES.md"
+  "README"
+  "docs/architecture/BACKEND-IMPLEMENTATION"
+  "docs/reference/COMPILER-CAPABILITIES"
+  "docs/history/ALPHA-1-RELEASE-NOTES"
 )
 
 # The machine-readable mirror (DOC-05): it follows the matrix, it does not
@@ -142,9 +145,15 @@ else
 fi
 
 # --- 2. Every over-claiming doc still points back at the matrix (DOC-04) -----
-for doc in "${BANNERED_DOCS[@]}"; do
-  if [ ! -f "$doc" ]; then
-    note "ERROR: bannered doc is missing: $doc"
+for stem in "${BANNERED_DOCS[@]}"; do
+  # Resolve the stem to whichever extension is present (.adoc preferred), so a
+  # .md<->.adoc migration does not trip the presence check.
+  doc=""
+  for ext in adoc md; do
+    if [ -f "$stem.$ext" ]; then doc="$stem.$ext"; break; fi
+  done
+  if [ -z "$doc" ]; then
+    note "ERROR: bannered doc is missing: $stem.{adoc,md}"
     note "       If it was intentionally removed, drop it from BANNERED_DOCS"
     note "       in this guard. Otherwise restore the file + its banner."
     fail=1

tools/doc-overclaims.allow

@@ -11,13 +11,13 @@
 docs/ROADMAP.adoc	* Production-ready toolchain
 docs/ROADMAP.adoc	|Production ready
 docs/STATE-2026-05-26.adoc	|*Phase 3 (Q3 2026)* |Production ready |*1.0* |requires Stage E (typed-wasm contract widening — cross-repo)
-docs/architecture/BACKEND-IMPLEMENTATION.md	> *production-ready*. One reference target (WASM); Deno-ESM/Node-CJS solid;
+docs/architecture/BACKEND-IMPLEMENTATION.adoc	_code-generators exist_, not _production-ready_. One reference target
 docs/bindings-roadmap.adoc	|Production backend default; pairs with SQLite for dev.
-docs/guides/WHAT-MAKES-IT-BRILLIANT.md	**Q: When will it be production-ready?**
-docs/history/ALPHA-1-RELEASE-NOTES.md	- Production-ready
-docs/history/ALPHA-1-RELEASE-NOTES.md	- Production-ready compiler
-docs/history/PHASE3-ASSESSMENT.md	- Row polymorphism is production-ready! ✅
-docs/history/PHASE3-COMPLETE.md	- **Production-ready** type system
-docs/history/PHASE3-SESSION-SUMMARY.md	- ✅ Row polymorphism is production-ready
-docs/history/PHASE3-SESSION-SUMMARY.md	| Row Polymorphism | 100% ✅ | Production ready |
-docs/standards/ROADMAP.md	This roadmap outlines the path from current state (lexer + parser) to a complete, production-ready language.
+docs/guides/WHAT-MAKES-IT-BRILLIANT.adoc	*Q: When will it be production-ready?* A: Aiming for compiler {plus}
+docs/history/ALPHA-1-RELEASE-NOTES.adoc	* Production-ready
+docs/history/ALPHA-1-RELEASE-NOTES.adoc	* Production-ready compiler
+docs/history/PHASE3-ASSESSMENT.adoc	*Good News:* - Row polymorphism is production-ready! ✅ - Effect
+docs/history/PHASE3-COMPLETE.adoc	*Lambda scope bug fixed* (bonus!) - *Production-ready* type system
+docs/history/PHASE3-SESSION-SUMMARY.adoc	Key accomplishments: - ✅ Row polymorphism is production-ready - ✅
+docs/history/PHASE3-SESSION-SUMMARY.adoc	|Row Polymorphism |100% ✅ |Production ready
+docs/standards/ROADMAP.adoc	to a complete, production-ready language.