feat(610): split the decision corpus into one YAML-frontmatter file per record

168 records -> docs/decisions/records/<area>/<topic>.md (163 active, 23 dirs) and
docs/decisions/archive/<area>/<topic>.md (5 archived). The filename IS the key,
so one-active-record-per-key becomes a filesystem property rather than a
validator check, and supersession becomes a `git mv`.

WHY: the monolith was a concurrency problem before an aesthetic one. A
3,900-line append target made parallel sessions collide -- PR #605 and PR #614
both hit append-vs-append conflicts during routine rebases, and hand-resolving
those inside the corpus is exactly the operation the rationale-rewrite guard
exists to police.

HOW IT IS VERIFIED: a ~170-file diff cannot be meaningfully read, so correctness
does not rest on reading it. The parser was taught BOTH formats first, so the
body-diff guard parses the old form at the merge-base and the new form at head --
the migration validates itself, no bypass. The proof is a field-level equivalence
harness: 168 records before and after, zero lost, zero gained, zero field
mismatches, zero rationale bodies differing. Reviewers should scrutinise the
harness; it is the actual evidence.

What measuring caught that reading would not have:

- ~500 lines sit OUTSIDE any record -- decisions.md's lifecycle schema and each
  topic file's preamble, mostly the only copy. Source files are kept and
  stripped, never deleted. They also cannot be filed per-area: topic files hold
  several areas and 4 of 23 areas span several files.
- Archive discovery was a non-recursive glob; after the split it found ZERO
  archived records, surfacing as four bogus "supersedes points to unknown key"
  errors rather than an obvious failure.
- ~32 live docs point into the corpus BY DATE, which the split dangles. Each
  stripped file now ends with a generated "Records formerly in this file" index,
  which also rescues the identical breadcrumbs in old issue comments.
- decisions.md's "In this file:" list was 97 same-file anchor bullets that the
  split makes WRONG, not merely stale. Dropped; the generated index replaces
  them with links that resolve.

The equivalence harness now runs against a checked-in FIXTURE, not the live
corpus. The earlier version migrated the real tree, which made it a one-shot:
the moment the migration landed there was nothing left to move and the tests
failed for reasons unrelated to the code. A fixture keeps them testing the
SCRIPT rather than the repo's current state.

Keys preserved verbatim, warts included: `sched` (12) and `scheduling` (1) remain
two directories for one concept. Renaming a key is not a move -- it changes
identity, breaks the equivalence proof, and invalidates MemPalace's per-key
drawers. Taxonomy normalisation is separate work.

refs #610

docs/decisions/records/ci/build-once-rejected.md

@@ -0,0 +1,30 @@
+---
+key: ci.build-once-rejected
+title: '2026-07-18 — CI build-once was measured and rejected; keep the #420 tree-skip'
+status: active
+since: '2026-07-18'
+supersedes: none
+superseded-by: none
+rule: 'CI build-once (a shared compile artifact across jobs) was implemented, measured, and rejected for a 40-85% wall-clock regression; keep the #420 cross-run tree-identity skip instead.'
+signals: 'CI build-once, artifact tar/transport cost, tree-identity skip · paths: `docs/ci-cd.md` → Cross-run tree-identity skip · issues: #420, #398, #455'
+mechanics: 'docs/ci-cd.md → Cross-run tree-identity skip; PR #455 measurement'
+---
+
+Build-once (a `compile` job producing a single artifact, consumed by `test`/`migrations`/
+`functional-e2e` via `--no-build`) was fully implemented and went **green on CI** (PR #455, run
+830), then **rejected on measurement**: it traded a ~12% slot-occupancy saving for a ~40–85%
+**per-run wall-clock regression**.
+
+- **Why it regressed.** The `bin`+`obj` artifact is 2.5 GB raw / 972 MB gz; tar alone costs ~82s CPU
+  plus ~180s transport, consuming most of the ~465s the shared compile was meant to save. Worse,
+  `compile` serializes **before** `migrations`' long ef-replay, which is pure DB work a shared build
+  cannot shorten — the bottleneck was never the redundant compiles.
+- **Incidental finding worth recording.** `actions/upload-artifact@v4` does not work on this Gitea
+  instance — it throws `GHESNotSupportedError`, because the `@actions/artifact` v2 client library
+  rejects any non-`github.com` host. `@v3` is required for any future artifact use here.
+- **Kept: the #420 cross-run tree-identity skip** (`docs/ci-cd.md` → Cross-run tree-identity skip).
+  It is independent of build-once — it only *skips* redundant work on identical-tree main pushes, at
+  zero wall-clock cost, rather than trying to share a build across jobs. Don't re-attempt build-once
+  unless the runner's artifact storage or network changes materially.
+
+Refs: #398 (closed), #420, PR #455.