Pipeline evidence + doc-sync hardening (#69 #70 #71 #72) → v6.4.0#73
Merged
Conversation
added 16 commits
June 3, 2026 12:43
Adopt existing <stem>-design-review.md/-plan-review.md convention (D1); complete D2 retro fix incl format-template:99 + Reads bullet; narrow Step 1e trigger + add PR-body accountability token (D6); reframe Resolution optional/ end-state with retro consumer (D5); clarify phase attribution (D4); dogfood caveat (D3). Commits the design-phase review report itself (dogfoods #69).
…st site, wired token consumer) N1 (Critical): replace ambiguous stem-derivation prose with deterministic design/plan rule + both worked examples (D1<->D2 path contract). N2: Step 1e gains a second edit site in finishing Autonomous Mode list. N3: wire Doc-reconciliation token to retro Step 5 missed-activation row. N4/N5: scalpel Reads edit; trigger reword.
…state, wire cross-refs CI)
…nner/append out (alignment PASS drift)
Contributor
There was a problem hiding this comment.
Pull request overview
This PR hardens the autonomous pipeline’s evidence trail and doc/implementation synchronization by making adversarial review outputs and activation evidence reliably consumable by the retrospective, and by adding a lightweight pre-PR doc reconciliation gate. It also adds a regression test + CI wiring to prevent these contracts from silently regressing, and bumps the kit version to v6.4.0.
Changes:
- Require
adversarial-design-reviewto produce a committed, deterministically-named review report with stable finding IDs (plus a plan-phase naming/identifier consistency check). - Update
post-merge-retrospectiveto treat.claude/autodev-state/in-progress.jsonlas the primary activation evidence source, gracefully degrading when absent, and to consume the Step 1e accountability token. - Add
finishing-a-development-branchStep 1e (Doc-Reconciliation Check) and a regression test (tests/pipeline-evidence-doc-sync.sh) wired into CI; bump manifests to 6.4.0.
Reviewed changes
Copilot reviewed 13 out of 13 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
tests/pipeline-evidence-doc-sync.sh |
Adds grep-based regression assertions to lock in the #69–#72 contracts. |
skills/adversarial-design-review/SKILL.md |
Mandates writing+committing the review report, adds stable IDs guidance, and adds a plan-phase identifier/naming-convention checklist row. |
skills/post-merge-retrospective/SKILL.md |
Switches activation scoring to use in-progress.jsonl as primary evidence; adds deterministic report-path derivation and Step 1e token consumption. |
skills/finishing-a-development-branch/SKILL.md |
Adds Step 1e Doc-Reconciliation gate and wires it into Autonomous Mode control flow. |
.github/workflows/skill-content-check.yml |
Runs the new regression test and tests/skill-cross-refs.sh in CI, and updates path filters accordingly. |
.claude-plugin/plugin.json |
Version bump to 6.4.0. |
.claude-plugin/marketplace.json |
Version bump to 6.4.0. |
.cursor-plugin/plugin.json |
Version bump to 6.4.0. |
docs/plans/2026-06-03-pipeline-evidence-doc-sync-design.md |
Adds the design doc for the change set (dogfooding the new workflow artifacts). |
docs/plans/2026-06-03-pipeline-evidence-doc-sync-design-review.md |
Adds the design-phase adversarial review report artifact. |
docs/plans/2026-06-03-pipeline-evidence-doc-sync.md |
Adds the implementation plan doc for the change set. |
docs/plans/2026-06-03-pipeline-evidence-doc-sync-plan-review.md |
Adds the plan-phase adversarial review report artifact. |
docs/plans/2026-06-03-pipeline-evidence-doc-sync.md.scope-lock |
Adds the scope-lock hash for the plan. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| - Any ADRs cited from the design or plan | ||
|
|
||
| 2. **Score each adversarial-review finding.** | ||
| Derive the report path by the **same deterministic rule as D1**: take the artifact filename, drop `.md`, then design → append `-review.md`, plan → append `-plan-review.md` (e.g. `…-doc-sync-design.md` → `…-doc-sync-design-review.md`; `2026-06-03-…-doc-sync.md` → `2026-06-03-…-doc-sync-plan-review.md`). Read the committed `…-design-review.md` / `…-plan-review.md` report(s). For each finding, use its stable ID; read the optional `Resolution` column as a scoring hint, **falling back to downstream evidence (code-review threads, CI) when blank or when the report is an old no-ID format**. If the report is absent → "no committed review report; reconstructed from revision history" (most pre-v6.4.0 features have none). |
| | brainstorming | yes | | | ||
| | adversarial-design-review (design) | yes | | | ||
| | adversarial-design-review (plan) | no | <why — e.g., manual override; deferred to alignment-check> | | ||
| | finishing Step 1e (doc-reconciliation) | yes/unverified | only when the diff touched docs/examples | |
Comment on lines
152
to
+159
| **Findings (Critical):** | ||
| - [class] [section/line]: <description>. Recommendation: <concrete fix>. | ||
| - `D1` [class] [section/line]: <description>. Recommendation: <concrete fix>. _Resolution: <optional — filled once at end-state: commit SHA / `accepted — reason` / `false-positive`; omit if open>._ | ||
|
|
||
| **Findings (Important):** | ||
| - [class] [section/line]: <description>. Recommendation: <concrete fix>. | ||
| - `D2` [class] [section/line]: <description>. Recommendation: <concrete fix>. _Resolution: <optional>._ | ||
|
|
||
| **Findings (Minor):** | ||
| - [class] [section/line]: <description>. Recommendation: <concrete fix>. | ||
| - `D3` [class] [section/line]: <description>. Recommendation: <concrete fix>. _Resolution: <optional>._ |
This was referenced Jun 3, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes four pipeline self-consistency issues with zero new skills/scripts and no heuristic scanner (per the user's anti-bloat / anti-trap / anti-circular-logic constraint). Net change ≈ +85 lines across 3 skills + a 63-line regression test + 3-manifest version bump.
adversarial-design-reviewnow mandates committing its findings report to the repo's existing<stem>-design-review.md/<stem>-plan-review.mdconvention, with stable finding IDs (D1/P1, sequential across all severities) + an optional end-stateResolution. This was a phantom dependency:post-merge-retrospectivealready read "adversarial-review reports committed in docs/plans/" but nothing wrote them — fixed by systematizing the ad-hoc practice (2 such files already existed) rather than inventing a new artifact.post-merge-retrospectivenow reads.claude/autodev-state/in-progress.jsonl(written by the shippedrecord-activityPostToolUse hook in any repo) as the primary activation source, demotes the kit-dev-onlytests/skill-activation-audit.sh, and degrades gracefully ("activation log unavailable", never "script does not exist"). Three edit sites fixed together (Step 5 + the## Missed skill activationsformat template + the**Reads:**bullet) so the broken instruction isn't re-embedded in every future retro.finishing-a-development-branchgains Step 1e: Doc-Reconciliation Check (pre-PR), wired into both the body and the Autonomous Mode list (so it fires in autonomous runs), checking committed docs/examples for (a) unlabeled forward-references → must bePlanned (PR #N)[alignment-check/scope-lock should scope-check committed docs & examples, not just plan tasks (forward-referencing docs slip through) #71] and (b) identifier drift vs the built code [Add a design-sync gate: design docs drift from implementation within locked scope (config keys, examples, formats) with no reconciliation step #72]. It's a checklist gate (not a scanner — the user's LIGHT choice) with a PR-bodyDoc-reconciliation:accountability token consumed by the retro. Plusadversarial-design-reviewgains one plan-phase Identifier / naming-convention match row (catch drift before code), explicitly distinct from the existing Config-validation row.Design / Plan / Adversarial review reports
docs/plans/2026-06-03-pipeline-evidence-doc-sync-design.mddocs/plans/2026-06-03-pipeline-evidence-doc-sync.mddocs/plans/2026-06-03-pipeline-evidence-doc-sync-design-review.mddocs/plans/2026-06-03-pipeline-evidence-doc-sync-plan-review.mdThese two committed
*-review.mdfiles are themselves the first dogfood of #69 (the new committed-report convention).Scope Manifest
PR Count: 1 · Tasks: 7 · Status: Locked 2026-06-03T17:10:43Z (sha256 6cb6ebd9…)
Changes (per task)
tests/pipeline-evidence-doc-sync.sh— grep-assertion regression guard for all 4 contracts (RED→GREEN TDD; verified genuinely RED onorigin/main).adversarial-design-review— committed-report mandate + deterministic path rule + stable IDs + Resolution +**Writes:**.adversarial-design-review— plan-phase naming-convention checklist row.post-merge-retrospective— jsonl-primary + 3 edit sites + Step-1e token consumer + back-compat.finishing-a-development-branch— Step 1e body + Autonomous Mode list item + accountability token..github/workflows/skill-content-check.yml— run the new test andskill-cross-refs.shin CI (the latter was local-only before).Verification
tests/pipeline-evidence-doc-sync.sh→Results: 0 failure(s)(revert-restore proven: each assertion goes RED when its contract is removed)tests/skill-content-grep.sh→ PASS (no host-token leaks) ·tests/skill-cross-refs.sh→ PASS (all refs resolve)tests/version-check.sh→ PASS (3 manifests = 6.4.0) ·tests/hook-contracts.sh→ PASS (no regression)plan-scope-check --verify-lock→ PASS (manifest untouched)a90b…→ the review commit.Doc-reconciliation: clean (Step 1e dogfooded on this PR — docs describe exactly this single PR's scope; identifiers match the built skills).
Generated by the autodev autonomous pipeline (brainstorming → adversarial design×3 → plan → adversarial plan → alignment → scope-lock → execute → adversarial code review). Copilot review is welcome but not blocking.
🤖 Generated with Claude Code