kairos-chain 3.61.1 → 3.63.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c28f2a9c842f5082b555797ea60f24d1435df8fcf80738556c88ffe341358468
-  data.tar.gz: 2cc735c1cf6c0cf86599c287cf1e9b095bad22e3de2f0e9055cf2bdf346a6b1c
+  metadata.gz: 96a6f46091b47614f44d089e6459dde7e8435cf7caa893c9b335b7359ad25394
+  data.tar.gz: 965451d952fb8fa61d3bfcc581241242de0f25eb492dddc56ebe913e4d07683b
 SHA512:
-  metadata.gz: 0e31b37f68c678274b1d3ea35dd5f9ce056971dc3e85f7697ee8203d8100a18261bb87ff425718f8bd7b4ce2c9d0b2c200b2b890aaf14dd6bd9e2367c722fc6e
-  data.tar.gz: af1ff31f3b3a8eac300e77b40d33fc8d41fa783be45c1cade91c88259dd9d456cd26a125b477127c869b324129357b0081d8bbfd12ee6b7a9bc73552eeb92f35
+  metadata.gz: 1c3812249637a923526a17b29cbbf34549f3f3039f51ae37941fa990ceec99611d68e88de280b39ff673aaf5f64e55be3133b079578b9f2887df8a6a6c81f102
+  data.tar.gz: a3688710826e2de6bd2d05ee24b0f9c6194e32e6d9dd8587c776b5e3bfa1b614549af903eee8a9bac295b5e0efc36c5e50537d5a54b8352acd40e05f2ef291c1

data/CHANGELOG.md

@@ -4,6 +4,85 @@ All notable changes to the `kairos-chain` gem will be documented in this file.
 
 This project follows [Semantic Versioning](https://semver.org/).
 
+## [3.63.0] - 2026-08-06
+
+### Changed
+
+- **`multi_llm_review` findings gain a weight axis, and a worker death no
+  longer discards completed seats** (SkillSet 0.9.1 → 0.10.0, frozen after one
+  review round; every deployment-grounded finding fixed in-round).
+
+  The severity of a finding said what kind of defect it was, never what it
+  costs: measured on a three-persona panel, 3 of one round's 7 P0s were
+  factually correct findings that cost nobody anything, landing at the same
+  severity as a defect that silently corrupts published output. The reviewer
+  prompt contract now requires a `[consequence: who is harmed, and how]`
+  clause on every P0, and aggregation records an unclaused P0 at P2 with the
+  stated severity and the demotion reason kept beside it. Only presence is
+  checked, mechanically; whether a stated consequence is real or trivial
+  stays the orchestrator's call. The clause is read before the byte bound
+  cuts the tail, the first non-empty clause counts, deduplication compares
+  issues without their clauses, and the demotion mark survives merging in
+  any member order. Reviewer prompts also stop naming the round number, and
+  the L1 workflow knowledge (v3.10.0) adds the matching orchestrator-side
+  rule: never tell a reviewer its finding count, the round number, or prior
+  verdicts — counts a reviewer performs select for finding-production over
+  finding-weight.
+
+  Separately, one stuck seat used to cost a whole round: a stale heartbeat
+  killed the detached worker and every completed seat's reply died with it,
+  because nothing left the worker's memory until all seats were done. The
+  worker now persists each seat's outcome the moment it is decided
+  (`partial_results.json`, atomic, single writer — including the
+  dispatcher's own deadline skips, so recovery cannot relabel a reached
+  seat as lost). The collect tool's crash and timeout branches recover the
+  completed seats, but only after the reaper CONFIRMS the worker's death: a
+  stale heartbeat is not a death certificate, and a record sealed over a
+  live worker would be permanently false once idempotent replay pins it.
+  Unconfirmed death falls back to the previous retryable total-loss report.
+  Seats the worker never reached enter the denominator as skip rows named
+  `worker_crashed_seat_lost`, and the payload carries `worker_failure`
+  naming the death and the recovered and lost seat labels.
+
+## [3.62.0] - 2026-08-06
+
+### Changed
+
+- **`project_orientation_report`'s checker stops rejecting reports that are fine.**
+  Three review rounds on a fixed panel, counting findings rather than approvals:
+  16, then 6, then 7. Nearly all of the first round's findings were reports a
+  reader would accept that the tool rejected, which is the failure mode that
+  matters for a lint — a missed defect costs an ugly figure, a false rejection
+  costs the author the report.
+
+  Adjacent text nodes were concatenated with nothing between them, so two table
+  cells holding `Ruby` and `3.4.10` produced a work-internal token `Ruby3` that
+  appeared nowhere in the document. Markup indentation counted toward the length
+  budget, so a report a browser renders as 3,556 characters was rejected at
+  6,226. `text-anchor` was not inherited from an enclosing group, so centred
+  labels were reported as overflowing by an amount that did not exist. A label
+  with no `x` inside a translated group was called unmeasurable although `x`
+  defaults to zero. Ambiguous-width characters had lost their width, so arrow and
+  box-drawing labels measured 45% short in the CJK font the template mandates.
+  Two-character Japanese labels did not count as a figure at all.
+
+  Reports now declare a worked prose example with `data-example="values"`,
+  satisfying invariant 8 the way the skill always promised. Body text is
+  NFKC-normalised before the token scan, so a full-width spelling — what a
+  Japanese input method emits by default — no longer slips past. Stylesheets are
+  parsed once, with comments stripped, for hiding, preformatting and absolute
+  font sizes alike.
+
+  Left open and stated as such: a rect is treated as a label's box when it
+  contains the label's whole width, which promotes a figure border to "the box"
+  and so disables overflow detection inside it. The opposite rule rejects
+  decorative highlight bands. Both are wrong in opposite directions; this ships
+  the one that errs toward passing. Choosing between them is a design question,
+  not a patch.
+
+  46 fixtures, 24 of which had never been added to git — a clone ran the suite
+  with them missing and no way to notice.
+
 ## [3.61.1] - 2026-08-06
 
 ### Fixed

data/lib/kairos_mcp/version.rb

@@ -1,4 +1,4 @@
 module KairosMcp
-  VERSION = "3.61.1"
+  VERSION = "3.63.0"
   CHANGELOG_URL = "https://github.com/masaomi/KairosChain_2026/blob/main/CHANGELOG.md"
 end

data/templates/knowledge/multi_llm_review_workflow/multi_llm_review_workflow.md

@@ -1,7 +1,7 @@
 ---
 name: multi_llm_review_workflow
 description: "Multi-LLM review methodology and execution — workflow pattern, CLI tooling, consensus analysis, Persona Assembly. Applicable to design, implementation, documentation, or any artifact."
-version: "3.8.0"
+version: "3.10.0"
 tags:
   - workflow
   - review
@@ -498,9 +498,58 @@ The user always has the final say.
     `open-question` or `defect` in the round's L2 record
     If no (a)/(b) blocking findings → proceed to next phase
     If any (a)/(b) finding          → repeat from [2] with revised artifact
+    (the revision obeys § Revision Discipline below)
     (c) findings are recorded as advisory; non-blocking
 ```
 
+## Revision Discipline (between rounds)
+
+> Evidence base: 35 recorded runs across 8 threads, 2026-08-03 → 08-06
+> (tokens in `.kairos/multi_llm_review/pending/`), two of which ran to
+> convergence. Same validation caveat as Step -1: a strong regularity in one
+> instance's logs, not yet reproduced elsewhere. Analysis record: L2
+> `mlr_p0_inflation_analysis_and_opus46_no_verdict_diagnosis_20260806`.
+
+The strongest predictor of round N+1's raw P0 count in those logs is whether
+the round-N revision **added mechanism** to the artifact — not the artifact's
+size, not reviewer strictness:
+
+- chain_history_erasure v0.5 added two invariants and a recount section
+  (draft 9.3k → 19.1k chars): raw P0 went 13 → 38, and 17 of the 38 targeted
+  the added or rewritten sections. v0.8, similar in size (18.6k) but authored
+  under an explicit "no new mechanism" rule, closed at 13 with one external
+  slot finding zero P0s.
+- mlr_evidence_fix R1's fix added four unrequested defensive mechanisms; R2
+  returned 41 findings, nearly all of them defects inside the additions. All
+  four mechanisms were later removed, each for a measured reason.
+- Deletions never generated findings: chain erasure v0.7 deleted three
+  mechanisms — zero new P0s against the deletions, confirmed in writing by
+  three slots.
+- Each thread converged within 1–2 rounds of switching to subtractive
+  revisions; neither converged while revisions were additive.
+
+Rules:
+
+1. **A revision closes findings by deletion, correction, or naming — never by
+   default-adding.** New mechanisms, invariants, sections, or defensive
+   layers do not enter a revision unprompted. If a finding appears to require
+   new mechanism, put the question to the operator ("finding X seems to need
+   mechanism Y — add, defer to backlog, or drop?") before drafting it in.
+   Explanatory prose is a lighter form of the same risk: a sentence added
+   only to justify a retreat became the sole blocking finding of the round
+   that followed it (mlr_evidence R6 — "adding an explanation creates a new
+   claim").
+2. **Prefer a revision author that is not the model whose additions are under
+   review.** This extends the existing separation principle — the deciding
+   context never authors what judges it — from verification to revision.
+   Opus 5 has a measured additive propensity (scope-broadening; the v0.5
+   explosion above), but the model is the pressure, not the cause: Fable 5
+   also added-and-broke (v0.6's fsync/realpath additions, v0.7's predicate 4
+   — a fatal genesis-rejecting defect) until the subtractive rule was
+   imposed, and Opus 5 converged mlr_evidence once its revisions became
+   subtractive (retreat + removal). Combine both levers: the subtractive
+   rule always, a different-model reviser when available.
+
 ## Review Types
 
 | Type | Focus | Reviewers See | Typical Use |
@@ -548,6 +597,19 @@ and philosophy-aligned finding has been answered is converged whether or not the
 numerator moved. Do not treat a reached ratio as sufficient on its own either:
 check what the approving replies actually said before counting them.
 
+**Count carryover and new (a)/(b) P0s separately; the machine-side signal of
+convergence is "new P0 = 0", not the APPROVE ratio.** Require each persona to
+state a closure verdict on its own prior-round P0s — closed / open /
+half-closed, with grounds. This format is validated live (chain erasure
+R6–R8) and is what makes the carryover/new split computable. A round whose
+(a)+(b) findings are all carryover with closure verdicts, and whose revision
+drew zero new P0s (observed without exception when the revision was
+subtractive — see § Revision Discipline), is a freeze candidate for the
+operator regardless of the numerator. Neither of the two 2026-08 threads
+ever reached its APPROVE ratio; both closed by (a)+(b) exhaustion + operator
+freeze declaration — the intended close described above, now with a
+measurable trigger.
+
 **Escalating raises the bar.** The rule is a ratio over the observers that
 counted, so adding reserve observers with `escalate: true` raises the number of
 agreements required. That is the intended cost of a wider panel, not a defect —
@@ -1027,6 +1089,26 @@ exclude it would be to exclude every honest terse approval with it. When a round
 reaches its ratio, read what the approving replies actually said before treating
 the ratio as convergence. That judgement is the human's and no rule replaces it.
 
+#### A no_verdict streak is a seat-environment signal, not a dead reviewer
+
+Before treating a slot as dead, read its `raw_text_excerpt` / `stated_text`
+in the pending record. Diagnosed live (2026-08-06, `claude_cli_opus4.6`,
+four consecutive no_verdict exclusions across one implementation-review
+thread): the CLI itself was healthy throughout. The subprocess seat runs
+sandboxed — no tools, empty working directory — so on **implementation**
+artifacts that cite file paths, the model attempted to read code before
+judging: two rounds opened with pseudo-tool-call markup, one opened with
+"the repository is not accessible" and stated its verdict header only
+further down, where the positional rule correctly refuses it. The same
+seat, in the same period, complied on **design** artifacts (verdict on
+line 1, counted every round). Remedies, in order: state in the subprocess
+prompt that the seat has no file access and must review the artifact text
+alone, marking unverifiable claims `[INFERRED]` (the grounding_rules block
+already licenses this); keep prompt rule #6 (full artifact inline) honest
+for implementation reviews; only then consider `--add-dir` with read-only
+tools, accepting CLAUDE.md contamination. A streak that survives those
+remedies is a real outage.
+
 #### Async/Parallel Collect Timing — Iron Rule
 
 When `delegation.parallel.default: true` (the v3.x default), Call 1 returns
@@ -1107,13 +1189,35 @@ Every review prompt MUST include these 7 items:
 1. **Output filename table** — so each reviewer knows where to save
 2. **Auto-execution commands** — ready-to-run CLI per reviewer
 3. **Review instructions** — what to focus on, what NOT to re-review
-4. **Review history** (R2+) — table of previous rounds and findings
+4. **Prior findings to verify** (R2+) — the findings the revision addresses,
+   so the reviewer can judge each closed / open / half-closed. Findings only:
+   no per-reviewer verdict history, no finding counts, no round tallies
+   (see § Reviewer incentive rule)
 5. **Context** — architecture summary for reviewers unfamiliar with codebase
 6. **Full artifact content inline** — reviewers may not have file access
 7. **Severity ratings + output format** — structured template for review output
 
 All prompt content MUST be in **English** for consistent parsing across LLM tools.
 
+### Reviewer incentive rule
+
+**Never tell a reviewer — subprocess or persona — that its finding count is
+compared across rounds, which round this is, or what verdicts were given
+before.** A reviewer told its count is watched treats the count as the
+deliverable, and that selects for finding-*production* over finding-*weight*
+(observed 2026-08-06: an orchestrator wrote "your finding count is compared
+across rounds" into persona prompts during the project_orientation_report
+loop; of the round's 7 P0s, 3 were factually correct findings that cost
+nobody anything). Convergence — carryover vs new, (a)+(b) exhaustion, the
+ratio — is measured by the orchestrator from the record, after the replies
+are in. The reviewer receives the artifact, the review criteria, and the
+prior findings it must verify. Nothing else about the loop's state.
+
+What this rule does NOT forbid: passing prior findings for closure
+verification (rule #4 — that is content, not score-keeping), and the
+carryover/new split in § Convergence Rules (that is orchestrator-side
+bookkeeping the reviewer never sees).
+
 ### XML Block Structure for Review Prompts
 
 Review prompts SHOULD use XML blocks to give LLMs explicit structural contracts.
@@ -1432,5 +1536,43 @@ Compression ratio: parallel agent raw → Assembly ≈ 2:1
   fence markers, character classes, digit ranges, word boundaries — and 18 of
   27 survived. Mutate the inside of a pattern, not only the pattern.
 
+- Revision discipline, new-P0 convergence signal, and no_verdict seat
+  diagnosis (v3.9.0, 2026-08-06): cross-thread analysis of 35 recorded runs
+  (8 threads, 2026-08-03 → 08-06) established that raw P0 growth tracks
+  additive revisions, not reviewer severity — every mechanism a revision
+  added became the next round's battleground, deletions drew zero new P0s
+  in every measured case, and both threads that converged did so within
+  1–2 rounds of switching to subtractive revisions (one under a Fable 5
+  reviser, one under the same Opus 5 orchestrator that had produced the
+  additive explosion). New § Revision Discipline encodes the subtractive
+  rule and the reviser-separation preference. § Convergence Rules gains the
+  carryover/new P0 split, with "new (a)+(b) P0 = 0" as the machine-side
+  freeze-candidate signal — neither thread ever reached its APPROVE ratio;
+  both closed by (a)+(b) exhaustion + operator freeze. § Substance and the
+  denominator gains the seat-environment diagnosis: `claude_cli_opus4.6`'s
+  four-round no_verdict streak was the sandboxed seat colliding with
+  implementation artifacts (pseudo-tool-calls, "repository not accessible"
+  preamble), not a dead reviewer — the same seat counted every round on
+  design artifacts in the same period. Analysis record: L2
+  `mlr_p0_inflation_analysis_and_opus46_no_verdict_diagnosis_20260806`
+
+- Reviewer incentive rule and the finding weight axis (v3.10.0, 2026-08-06):
+  § Prompt Generation Rules gains the Reviewer incentive rule — reviewer
+  prompts never mention finding counts, round numbers, or prior verdicts;
+  convergence is measured orchestrator-side from the record, and prior
+  findings are passed for closure verification only (rule #4 reworded
+  accordingly, from "review history table" to "prior findings to verify").
+  Motivating observation: an orchestrator told personas their counts were
+  compared across rounds, and 3 of the round's 7 P0s were factually correct
+  findings that cost nobody anything. In the same change the SkillSet
+  (0.10.0) adds the weight axis mechanically: the prompt contract requires a
+  `[consequence: who is harmed, and how]` clause on every P0, and
+  aggregation records a P0 without one at P2, keeping the stated severity
+  and the demotion reason beside it (`severity_stated` /
+  `severity_demoted: consequence_missing`). Presence is checked
+  mechanically; whether a stated consequence is real or trivial stays the
+  orchestrator's call, per the (a)/(b)/(c) discipline. Handoff record: L2
+  `handoff_mlr_finding_weight_axis_and_reviewer_incentive_20260806`
+
 **Key insight**: Design reviews and implementation reviews find
 **categorically different bugs**. Both phases are necessary.

data/templates/knowledge/project_orientation_report/project_orientation_report.md

@@ -153,13 +153,19 @@ SVG の制約。守らないと壊れる。
 6. **人が読む。** 通らなかった節が分かったら、**レポートではなくこの skill を直す**。
    条件 7・8・9・10 はすべて、この段の判定から出た。
 
-## レポート側が守る 2 つの約束ごと
+## レポート側が守る 3 つの約束ごと
 
-機械が読むための印を 2 つ、レポートの HTML に入れる。
+機械が読むための印を 3 つ、レポートの HTML に入れる。
 
 **図表の無い節は、見出しで宣言する。** `<h2 data-visual="none">決まっていないこと</h2>`。
 宣言の無い節は図か表を要求される。位置で免除しない——節が 1 つ増えただけで免除が
-隣の節へずれるため。
+隣の節へずれるため。宣言できるのは 3 節まで（表紙、「決まっていないこと」、
+食い違いが無いときの「記録の食い違い」で使い切る想定）。
+
+**図表の代わりに実際の値の例で満たす節は、その要素に印を付ける。**
+`<p data-example="values">入力 X を送ると出力は Y だった。</p>`。
+条件 8 は「図・表・実際の値による具体例のいずれか」を認めており、この印が
+「値の入った例」を機械に伝える。印の無いただの散文は数えない。
 
 > この宣言は v0.2 で入った。**それ以前に書かれたレポートは、表紙と「決まっていないこと」の
 > 見出しに属性を 1 つずつ足すまで落ちる。** 読み手に見えるものは変わらないので、
@@ -170,7 +176,15 @@ SVG の制約。守らないと壊れる。
 `<meta name="orientation-allowed-tokens" content="A4, Q3, gpt-5">`。
 検出はわざと広くしてあるので、紙の大きさやモデル名のような無害な語も引っかかる。
 それを消すには宣言を書く。**宣言を書くこと自体が規律**で、スクリプトの中の許可リストを
-太らせるより安全である。組み込みの免除は層の名前 3 語だけ。
+太らせるより安全である。組み込みの免除は層の名前 3 語だけ。宣言できるのは 16 語まで——
+節の免除と同じで、宣言で要求を丸ごと消す形は認めない。この上限は、条件 7 を生んだ
+「符牒 13 種で読めなかったレポート」への余白として選んだ数ではない。あの 13 種は
+**宣言されずに**本文へ出た符牒で、上限はそこでは発火しない（自分の符牒を全部宣言してしまう形は、
+適用範囲の表にあるとおり、意図しなければ起きない容認済みの抜け道である）。一方で、実在する
+技術の段落が SHA-256 や Ed25519 のような公開語彙を 13 語、正当に宣言する必要があった。
+上限が止めるのは宣言で要求を丸ごと消す形だけなので、正当な宣言の実測最大 13 語を
+落とさない粗い天井として 16 にしてある。`v0.9` や `TLS-1.3` のような
+語は書いたとおりに宣言すればよい（部分文字列に割って宣言し直す必要は無い）。
 
 ## 検収
 
@@ -179,13 +193,13 @@ SVG の制約。守らないと壊れる。
 
 | 検査 | 合格条件 |
 |---|---|
-| 未記入 | 《》で囲まれた placeholder が 0 個。未記入のテンプレートは通らない |
-| 符牒 | 見えている文に、宣言されていない符牒が 0 個 |
-| 整形済みの塊 | `<pre>` と `white-space: pre` が本体に 0 個 |
-| 分量 | 見えている文が 6000 文字以内（`--max-body-chars` で変更可） |
+| 未記入 | 《》で囲まれた placeholder が 0 個。未記入のテンプレートは通らない。SVG の `<title>` も読み上げられるので対象 |
+| 符牒 | 見えている文に、宣言されていない符牒が 0 個。全角で書いた符牒（Ｆ１）も NFKC 正規化で同じ符牒として数える。宣言は 16 語まで |
+| 整形済みの塊 | `<pre>` と `white-space: pre` が本体に 0 個。付録や印刷だけに効く stylesheet の規則は数えない |
+| 分量 | 見えている文が 6000 文字以内（`--max-body-chars` で変更可）。markup の字下げは数えない |
 | 節 | 8 節以上 |
-| 視覚要素 | 宣言で免除されていない全節に、**中身のある**図か表がある（空の表は不可） |
-| 図の収まり | すべての行が、囲んでいる `<rect>` に収まる。箱が無ければ図の枠に収まる |
+| 視覚要素 | 宣言で免除されていない全節に、**中身のある**図か表、または `data-example="values"` を付けた値の例がある（空の表は不可）。図として数える label は 2 文字以上。免除宣言は 3 節まで |
+| 図の収まり | すべての行が、囲んでいる `<rect>` に収まる。箱とみなすのは行の全幅を含む最小の `<rect>` で、全幅を含む `<rect>` が無いときだけ開始点を含む最小の `<rect>` と照合する（強調の帯を箱と誤認しないため）。`<rect>` が無ければ図の枠に収まる |
 
 「見えている内容」の定義がこの検収の要である。折りたたまれていない `<details open>` の中も、
 `<svg>` の中の `<text>` も、読み手には見えているので本文に数える。`<script>` `<style>`
@@ -200,7 +214,7 @@ SVG の制約。守らないと壊れる。
 |---|---|
 | ASCII 図、未記入の欄、箱から溢れる label | 別紙の CSS が `content:` で描く文字 |
 | 測り方が拾えない書き方の `translate` | `<iframe srcdoc>` の中身 |
-| `em` のような相対単位の font-size | 全角や幅ゼロ文字を混ぜた符牒 |
+| `em` のような相対単位の font-size、全角で書いた符牒 | 幅ゼロ文字を混ぜた符牒 |
 | 全部の節を免除にしてしまう | 禁止すべき符牒を自分で許可宣言する |
 | 見出しを `display:none` で隠す | 入れ子の `<svg>` を親の外へ置く |
 | 折りたたんだ付録の `<summary>` に本文を書く | `textLength` で描画幅を上書きする |
@@ -212,8 +226,9 @@ SVG の制約。守らないと壊れる。
 
 ### 検収が測っていないもの
 
-不変条件は 10 個あり、**機械が見ているのは 3 個だけ**（符牒、視覚要素と分量、図の収まり）。
-残る 7 個は人が守るしかない。
+不変条件は 10 個あり、**機械が見ているのは 3 個だけ**（7 符牒、8 視覚要素、9 分量）。
+図の収まりも機械が測るが、あれは「図の作り方」の規約の検査であって、10 個の不変条件の
+うちの 1 つではない。残る 7 個は人が守るしかない。
 
 | 未検査の条件 | なぜ機械で測れないか |
 |---|---|
@@ -230,10 +245,14 @@ SVG の制約。守らないと壊れる。
 
 ## テスト
 
-`test/test_check_report.py` が `test/fixtures/` の 14 件を走らせる。標準ライブラリだけで動く。
+`test/test_check_report.py` が `test/fixtures/` の 46 件を走らせる。標準ライブラリだけで動く。
 
-fixture は **2 方向**ある。10 件は「読めないのに通ってしまった形」で、すべて外部の評価者が
-実際に実行して示した抜け道である。4 件は「読めるのに落とされた形」。
+fixture は **2 方向**ある。25 件は「読めないのに通ってしまった形」、21 件は「読めるのに
+落とされた形」。落ちる側のうち 8 件は外部の評価者が実際に実行して示した抜け道で
+（最初の 10 件のうち `bad_missing_section.html` と `bad_comment_details.html` の 2 件は
+初版でも「節」で落ちていたので抜け道ではない）、
+残りはその後のレビューで見つかった、書き手が普通に書いていて踏む事故の形と、
+修正の後退を防ぐ検査である。
 **塞ぐだけなら全部不合格にすれば達成できる**ので、両方向を同時に測る。
 どの検査が捕まえるべきかも fixture ごとに宣言してあり、別の理由で落ちた場合はテストが赤くなる。
 
@@ -253,5 +272,8 @@ KairosChain の機能ではない。KairosChain 単体で動かす場合は、
 
 **実装は 1 度 multi-LLM review に落ちている。** 4 者中 0 者が承認、深刻な指摘 14 件。
 初版の検収スクリプトはタグを見ていたので、内容を除外領域へ移すだけで全部すり抜けられた。
-評価者が実行して示した抜け道が、いま `test/fixtures/` の `bad_*` 10 件になっている。
+評価者が実行して示した抜け道が、いま `test/fixtures/` の `bad_*` のうち最初の 10 件中
+8 件になっている（`bad_missing_section.html` と `bad_comment_details.html` は初版でも
+「節」で落ちていたので抜け道ではない。残りの `bad_*` も後のレビューで加わった事故の形と
+後退防止で、評価者由来ではない）。
 経緯は `references/worked_example.md`。

data/templates/knowledge/project_orientation_report/references/worked_example.md

@@ -55,9 +55,17 @@
 
 ## fixture — 破られた形がそのまま検査になっている
 
-`test/fixtures/` の 14 件。`test/test_check_report.py` が走らせる。
-
-**通ってはいけない 10 件**（すべて評価者が実行して示した抜け道）
+`test/fixtures/` の 46 件。`test/test_check_report.py` が走らせる。
+内訳は四世代ある。初版で破られた形（下の 2 表）、レビュー 2 周目で入った
+「書き手が普通に書いていて踏む事故の形」8 件、3 周目のレビュー指摘の
+修正に付けた後退防止の検査、そして 4 周目（普通に書いたレポートに誤った答えを
+返す形の修正）で加わった検査。どの fixture をどの検査が捕まえるべきかの宣言は
+`test_check_report.py` にある。
+
+**落ちる側の fixture のうち、初版由来の 10 件。うち抜け道は 8 件**（評価者が実行して
+示したのは 8 件。※印の 2 件——`bad_missing_section.html` と `bad_comment_details.html`——は
+初版（9c3eb0a）でもどちらも「節」の検査で落ちており、抜け道ではなく最初からある検査の
+確認である。それ以降に追加された落ちる側の fixture も評価者由来ではない）
 
 | fixture | 何を再現しているか | 捕まえる検査 |
 |---|---|---|
@@ -68,11 +76,12 @@
 | `bad_svg_overflow_rect.html` | 図の枠には収まるが箱から溢れる label | 図の収まり |
 | `bad_svg_transform.html` | `translate` で画面外へ出る label | 図の収まり |
 | `bad_token_forms.html` | `f1` `c-1` `step-3` `#12` のような形 | 符牒 |
-| `bad_comment_details.html` | コメントの中の `<details` で本文を切る | 整形済みの塊 |
+| `bad_comment_details.html` ※抜け道ではない | コメントの中の `<details` で本文を切る | 整形済みの塊 |
 | `bad_whitespace_pre.html` | `<pre>` を使わずに ASCII 図を出す | 整形済みの塊 |
-| `bad_missing_section.html` | 節が 1 つ足りない | 節 |
+| `bad_missing_section.html` ※抜け道ではない | 節が 1 つ足りない | 節 |
 
-**落としてはいけない 4 件**
+**落としてはいけない fixture のうち、初版からの 4 件**（その後、レビュー指摘で判明した
+「正しく書いたのに落とされる形」を通す fixture が 17 件加わり、通る側は計 21 件）
 
 | fixture | なぜ通らねばならないか |
 |---|---|