Collapse the done entries, and re-run the budget on a body edit

[laidrivm]

Reduces each completed queue entry in PLAN.md to its outcome and its archive path. The always-on set goes 495 lines to 479 against the ~500 the context-budget capability sets.

Closes no acceptance criterion. It applies one that already stands — PLAN.md holds the open queue and the standing constraints, which says a completed entry "SHALL NOT carry the narrative of how it was implemented" — rather than changing anything the spec says.

Also carries a fix the stack turned up: the diff-budget workflow named no activity types, so it listened for opened, synchronize and reopened alone. The oversize: override lives in the pull request body, so a marker added after the last push was never read — the check kept reporting a verdict on a body that had since changed, and only an empty commit could clear it. #87 was cleared by accident, a merge commit happening to supply the synchronize.

change-slicing requires CI to fail unless the body carries the marker, which is a statement about the body as it stands, so this is the implementation catching up with the spec rather than a change to it. The trigger is pinned by a test that also asserts the three defaults survive — naming types at all is what stops them applying, so fixing one and silently losing the others is the easy mistake.

Every fact dropped was read at the place that owns it before it was dropped, which is what the spec asks for instead of assuming the archive holds it: the boundary at exactly 800 in change-slicing's living spec, the two floors at their FLOOR constants and the reason each last moved beside them, the refilling settings.local.json in the agent-permissions spec, the equivalent mutants in mutation-floor's archived design and its living spec.

Two did not survive that test and did not go:

• The ask entries confirmed live in a session started after the merge appear in no archived change. An archive is not edited to receive a fact found later, so it stays as that entry's outcome.
• The skills-lock.json patch skill-provenance drafted is work nobody has done. It moves to the open queue, where it belonged.

To check by hand: nothing runs over PLAN.md. The reviewer's check is that no dropped sentence names a fact the tree does not carry elsewhere.

Summary by CodeRabbit

• New Features

  • Pull request checks now run when a pull request description is edited, in addition to existing activity triggers.

• Documentation

  • Updated project tracking to consolidate completed work, add archive references, and clearly list an outstanding lockfile update.

• Tests

  • Added coverage confirming checks run for pull request description edits and existing pull request events.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: ae8db349-02aa-4705-9b05-4a19ac5e98b9

📥 Commits

Reviewing files that changed from the base of the PR and between b1342b0 and 60235ab.

📒 Files selected for processing (3)

• .github/workflows/diff-budget.yml
• PLAN.md
• scripts/diff-budget.test.ts

---

📝 Walkthrough

Walkthrough

The diff-budget workflow now runs for edited pull requests. A regression test verifies the complete event set. PLAN.md updates completed-work records and adds a pending skills-lock.json item.

Changes

Workflow and plan updates

Layer / File(s)	Summary
Diff-budget edited trigger .github/workflows/diff-budget.yml, scripts/diff-budget.test.ts	The workflow now handles edited pull request events. The test verifies edited, opened, synchronize, and reopened.
Plan queue records PLAN.md	The Done queue entries were condensed and reorganized. The Open queue now lists the unapplied skills-lock.json patch.

Estimated code review effort: 2 (Simple) | ~10 minutes

Mergeability Score: 🟡 Moderate · up to 60235

The PR makes the budget check reevaluate pull-request body edits and condenses completed planning entries, but the added regression test appears syntactically invalid because the YAML parse expression is repeated without separators, so merge should wait until that test is corrected. The plan rewrite also needs bounded cleanup to verify the retained historical claim and use complete archive paths.

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes both main changes: condensing completed plan entries and rerunning the budget check after body edits.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch chore/collapse-plan-done

---

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