docs(scripts): document sanctioned in-place CHANGELOG body corrections

[kyle-sexton]

Summary

Documents the #2388 decision in scripts/check-changelog-parity.sh: declared in-place corrections inside an already-released ## [<v>] section are sanctioned when the correcting PR names each edit in its body and in the new release entry.

--check-preserved continues to enforce heading survival; body fidelity is review discipline, not an automated gate.

Decision brief (#2388)

RECOMMENDED: Option 2 — declared in-place correction is sanctioned with explicit naming in the correcting PR (the de facto behavior #2312 followed). Option 1 (append-only errata only) is not adopted.

Closes #2388.

Related

• Refs fix(source-control): D6 resolves the push remote instead of hardcoding origin #2312 (the de facto declared-correction behavior this documents).

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

[claude]

Claude finished @kyle-sexton's task in 20s —— View job

---

Security Review

• Fetch PR diff and read changed files
• Analyze for security issues (logic flaws, authz gaps, injection, secrets handling, dangerous Actions patterns)
• Post findings

No security issues found.

This PR changes exactly one file — scripts/check-changelog-parity.sh — and the entire diff is a 6-line addition to the existing header comment block (lines 27–33) documenting the sanctioned in-place CHANGELOG-correction policy. No executable code, control flow, argument handling, or shell logic is touched, and there are no workflow, permissions, or dependency-pin changes in scope. There is no security-relevant surface to review here.

[claude]

Claude finished @kyle-sexton's task in 1m 50s —— View job

---

Review complete ✅

• Gather context (PR diff, CLAUDE.md, REVIEW.md)
• Review changed file for correctness and alignment with repo guidelines
• Post findings

Scope: scripts/check-changelog-parity.sh (+6/-0, comment-only). Per REVIEW.md this PR is reviewed under the code-review lane (correctness, design, conventions, docs); no security-relevant surface here.

Summary

This adds a documentation-only comment block after the --check-preserved usage description, recording the #2388 decision: declared in-place corrections to an already-released ## [<v>] section are sanctioned when the correcting PR names each edit in both its body and the new release entry, and that --check-preserved remains a headings-only mechanical gate (body fidelity stays a review-time discipline, not automated).

I verified this claim against the actual implementation:

• missing_preserved_headings() (line 246) and the --check-preserved mode block (lines 462–570+) operate purely on ## [<v>] heading presence/absence — they never inspect section body content. The new comment's characterization of the gate's scope is accurate.
• The bare (#2388) issue reference style matches existing precedent in the same file (e.g. line 645's (#2130 ...)), so it's consistent with in-file convention rather than introducing a new style alongside the fully-qualified claude-code-plugins#NNNN form used elsewhere (lines 46, 68, 85).
• Diff is exactly 6 added comment lines, no code/logic change, matching the PR's stated scope — no test coverage gap since no behavior changed.

No correctness, design, or convention issues found. Nothing to flag as blocking or as a suggestion.

Note: I could not verify the referenced #2388 decision brief or the #2312 precedent it cites against their source issues/PRs directly (gh CLI calls required approval not available in this session) — I'm relying on the PR body's description and validating internal consistency with the code, which checks out.

[github-actions]

Claude has reviewed this PR 1 time. The lane skips further automatic reviews after 5; deleting this comment resets the count.