Board hygiene for merged #971

[AdaWorldAPI]

Discharges the post-merge obligation from CLAUDE.md's board-hygiene rule: a merged PR owes a PR_ARC_INVENTORY entry and a LATEST_STATE record. #971 merged as 2cbe62d (head d627f5c).

Diff is 172 insertions, 0 deletions — the structural append-only audit for these two files.

PR_ARC_INVENTORY.md (prepend)

The full Added / Locked / Deferred / Docs / Confidence entry, covering all four stages: the maturity gate + four kernel carves, the corrected Stage-2.5 census, Stage-2.6a V3 invariance, and the CE64 ⇄ V3 losslessness gate.

The entry is long because six of its items are corrections the arc made to its own record, and each is worth more than the result it corrected:

1. the audit example was pinning the defect it tested;
2. the consumer filter tested a predicate that could not matter — which inverted the published headline from 0/5,760 to 1,098/5,760;
3. the report hardcoded its own conclusion as a literal;
4. a coarse label certified a component it never looked at;
5. the Stage-2.6 brief's premise was false and was re-scoped;
6. InferenceType was rewriting half the mantissa state space on every conversion — and the Stage-2.6a harness could not have caught it, because it only ever carries a surviving state.

Confidence is graded per claim rather than for the PR. Notably MEDIUM on moves_confidence()'s 34 hand-written declarations: they are pinned two-sided against the probe matrix, but that matrix is a sample — a kernel with a confidence branch no probe reaches would be declared false and the pin would agree with it. Flagged to reviewers at the time; recorded here rather than closed.

LATEST_STATE.md (prepend)

The merged-PR record — a four-row stage index over the five dated sections already in the file, plus the one finding a future session should not have to rediscover (InferenceType is lossy on 8 of 16 states, carry the raw nibble; the tail fields cross as raw ordinals, and preservation is not a provenance upgrade), plus the four items left open as Stage-3 inputs.

Not added to the "Recently Shipped PRs" table: that table stalled at #946 and carries its own gap note ruling that recent PRs are held by the dated sections + PR_ARC_INVENTORY. Following the live convention rather than reviving a stale one.

Termination

This PR's entire content is board hygiene for a prior PR. Per CLAUDE.md's termination clause it therefore generates none of the merged-PR row's obligations — no arc entry and no LATEST_STATE update are owed for it. It is discharged by the entries it wrote. Do not open a follow-up.

---

Generated by Claude Code

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 84077bb4-6e02-4efe-9368-486652e4aa63

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}