docs(autonomy): record every un-tracked D1 deferral as a durable trigger row

[kyle-sexton]

Summary

D1 sweep for #353: the WP1–WP7 PLAN out-of-package notes (mined from the merged PR bodies #333/#343/#356/#372/#377/#600/#676) contained 31 deferred notes; every note that named no tracking issue and had no contract-recorded trigger now has a durable home:

• README roadmap gains four trigger-gated rows: fleet guardrail materializations (WP5), fleet routine stand-up + existing-scheduler reconciliation (WP6), vendor-binding capability templates (WP6), cost enforcement / hard spend caps (WP5, 3→4 trigger).
• Trigger register gains the second-binding-consumer cross-repo drift check (WP1).
• reference/return-accounting.md records the per-work-class precision-graduation deferral beside its band-stability rule (WP3).
• CHANGELOG.md created for the autonomy plugin (first entry; 0.1.0–0.7.0 history stays in the WP PR bodies) + version bump to 0.7.1 — starts the CHANGELOG-parity posture ci: orphaned-fixture gate + CHANGELOG-parity gate (every fixtures/evals file consumed by a grader; every plugin.json bump requires a CHANGELOG entry) #663 gates on.

All other mined notes were already covered: tracked issues (#351, #352, #694–#697), contract-recorded triggers (telemetry immutable-ID, return-accounting expansion/revisit), delivered sibling WPs, resolved /architect questions, or the user-held dormant triggers (runner build T4, L3 backend, merge serialization, mid-run interrupt, org-enablement, cross-team). Full disposition table lands as a comment on #353 at close.

Documentation only — no contract semantics change.

Related

• docs(topics): prune delivered AI-ladder WP slices at effort end #685 (effort-end prune that moved the PLAN records into the merged PR bodies)
• ci: orphaned-fixture gate + CHANGELOG-parity gate (every fixtures/evals file consumed by a grader; every plugin.json bump requires a CHANGELOG entry) #663 (CHANGELOG-parity gate this PR's new CHANGELOG.md starts satisfying)
• WP delivery PRs mined: feat(autonomy): autonomy plugin — category, topology contracts, guided-setup v0 (WP1) #333, feat(autonomy): telemetry contract + setup slice, conformance check, live-proven demo (WP2) #343, feat(autonomy): return-accounting convention, capture slice, live-proven demo (WP3) #356, feat(autonomy): trigger-dispatch package - contract, setup slice, envelope check (WP4) #372, feat(autonomy): guardrail matrix, security binding, setup slice (WP5) #377, feat(autonomy): standing-routine catalog, v1 leaves, setup routine slice (WP6) #600, docs(autonomy): autonomous-runner design pack (WP7) #676

Closes #353

🤖 Generated with Claude Code

[claude]

Claude finished @kyle-sexton's task in 1m 13s —— View job

---

PR Review — docs(autonomy): record every un-tracked D1 deferral as a durable trigger row

• Gather context & read changed files
• Review plugin.json version bump
• Review CHANGELOG.md (new file)
• Review README.md additions
• Review reference/return-accounting.md additions
• Post final findings

---

Scope: documentation-only — no contract semantics change. No security surface, no code execution, no secrets, no egress. All four files reviewed against the diff.

plugin.json

0.7.0 → 0.7.1 — correct. Patch bump for documentation-only additions is the right SemVer signal; no API or contract surface changed. File is otherwise valid.

CHANGELOG.md (new file)

Lines 1–19

The file cites Keep a Changelog 1.1.0 as its format, but ## [0.7.1] is missing the date component that format mandates: ## [version] - YYYY-MM-DD. Minor omission — the entry is otherwise well-structured. Keep a Changelog also recommends reference links at the bottom of the file (e.g., [0.7.1]: https://github.com/…/compare/v0.7.0...v0.7.1), though those are commonly omitted on first entries.

Suggestion (non-blocking): Add the release date to the version header:

## [0.7.1] - 2026-07-20

Fix this →

README.md — roadmap additions

Lines 67–71

The four new roadmap rows are consistent in style with the existing fleet-adapter row. Trigger descriptions are appropriately gated on post-package graduation or a named transition. One wording note: 3→4 transition work begins (cost enforcement row) is compact shorthand — it will read clearly to anyone familiar with the adoption-ladder vocabulary, but a brief parenthetical (e.g., adoption-ladder stage 3→4) would make it self-contained without the plugin's broader context. Non-blocking.

Line 79 — trigger register addition ("A second repository consumes the security binding → Stand up a mechanical cross-repo binding drift check") is actionable and consistent with the two existing trigger rows.

reference/return-accounting.md

Lines 59–60

Placement is correct — the precision-graduation deferral sits immediately after the contract-stability rule it qualifies. The trigger condition ("aggregate data proving a class needs finer resolution") is concrete and matches the conditional-not-permanent pattern used for the agent/LLM layer revisit trigger on line 18. No issues.

---

Summary

Clean, low-risk documentation PR. One actionable suggestion: add the ISO date to the ## [0.7.1] CHANGELOG header to honor the Keep a Changelog format the file cites. Everything else is correct and well-formed.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c02ccf2e0c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[claude]

Claude finished @kyle-sexton's task in 1m 12s —— View job

---

PR Review — docs(autonomy): record every un-tracked D1 deferral as a durable trigger row

• Gather context & identify new commit
• Read diff of latest commit (dd24700)
• Review return-accounting.md changes (class-evidence path addition)
• Verify CHANGELOG.md date status
• Post final findings

---

Scope: This pass covers the follow-up commit dd24700, which responds to the Codex P2 finding on reference/return-accounting.md. The earlier review of plugin.json, CHANGELOG.md, and README.md (commit c02ccf2) stands unchanged.

dd24700 — reference/return-accounting.md (class-evidence path)

Lines 59–65

The Codex P2 concern was that the precision-graduation trigger cited "aggregate data proving a class needs finer resolution" but provided no path to obtain per-class effort-band aggregates — since the record stores no class field and the telemetry contract only defined the work-item URL as a join key.

The fix names the full evidence path without adding a record field:

1. Join path: each record → its work item via the telemetry contract's join attribute
2. Primary source: the item's admission-time class from the governed queue's protected admission data (the surface that stamped and verified the class at admission)
3. Fallback: re-derivation through the security-surface classification rules, explicitly marked as a "stated approximation" for cases where queue history is not retained

This is the correct fix for the finding. The approach keeps the two-field record minimality intact (the contract-stable invariant the prior text already established) while making the trigger evaluable: an adopter accumulating records can now join backward to per-class effort-band distributions through the specified path, with a documented approximation for the incomplete-history case. The "stated approximation" qualifier is important — it accurately communicates that fallback re-derivation reflects current-epoch class, not necessarily admission-time class.

Prose density is high but consistent with the surrounding document style. No issues.

CHANGELOG.md — date still absent

The ## [0.7.1] header in CHANGELOG.md still lacks the ISO date mandated by the Keep a Changelog 1.1.0 format the file cites. This was flagged as non-blocking in the first review pass; it remains the one open item.

Fix this →

---

Summary

dd24700 correctly resolves the Codex P2 finding: the precision-graduation trigger is now evaluable through a documented join path, and record minimality is preserved. One non-blocking item remains: add the ISO date to ## [0.7.1] in CHANGELOG.md. Otherwise ready to merge.