chore(ai-slop): disable rule-emoji-formatting here, with the measurement

[kyle-sexton]

No linked issue

Summary

All 21 rule-emoji-formatting findings against this repo are semantic markers; none is decoration. The rule is doing its job — this repo simply uses these glyphs as vocabulary rather than ornament, which is a house-style call the config layer exists to record.

Measured 2026-08-20 over the 1214-file corpus, fence-stripped the way the detector does:

Class	Count	Where
Wrong/right teaching pairs (❌/✅)	7	docs-hygiene citation-form.md, planning interview/loop.md
Coaching pairs (✗/✓)	4	pat-pattison coaching-protocol.md
Warning headings (⚠)	9	kindle-dedrm README, songwriting CHANGELOG, book-references.md
Severity-legend marker (🟣)	1	REVIEW.md, sibling to the 🔴/🟡 in its own table

Fix

rule-emoji-formatting is added to disabled_rules in .claude/ai-slop.json, with the measurement and the class breakdown recorded under the single supported _comment key — the same treatment and the same rationale shape as rule-em-dash and rule-curly-artifacts already carry. The shipped detector default stays neutral; this is the consuming repo's own taste.

Why not in-file block markers. That was the first choice, and it was rejected once its cost was measured rather than assumed. The marked files span four plugins (docs-hygiene, kindle-dedrm, planning, songwriting), and changelog-parity-gate treats any edit under plugins/<name>/ as published-version reuse — the failure mode that blocked #3054 until a bump was added. Silencing cosmetic audit noise would therefore have published four plugin releases. This change touches one repo-root file and bumps nothing.

Why not a per-path exemption. It is the instrument that actually fits all three disabled entries, but the config layer supports per-rule path allowlisting only for rule-em-dash today. Generalizing it is a plugin change with its own design and test surface, and does not belong in a consumer-config diff. The _comment records that so the next reader does not re-derive it.

Verification

• detect.sh --show-config → disabled_rules=rule-em-dash rule-curly-artifacts rule-emoji-formatting.
• Scoped run over the affected files: rule-emoji-formatting findings=0 declined=0 disabled=1, Summary total: 0 findings.
• .claude/ai-slop.json still carries exactly three keys — _comment, disabled_rules, excluded_paths — so no unknown-key drift against the ai-slop:setup contract (the defect caught by review on chore(ai-slop): honor the repo's standing punctuation-glyph ruling in config #3054).
• check-changelog-parity.sh both --check and --check-bump origin/main — pass. No plugin file is touched, so no version bump is owed.
• detect.test.sh — 92 cases pass (unchanged by this diff; run to confirm the config change does not leak into the suite's pinned-empty cascade).
• editorconfig-checker, typos — clean.

Related

• Refs chore(ai-slop): honor the repo's standing punctuation-glyph ruling in config #3054 — the rule-curly-artifacts entry and the _comment single-key contract this follows.
• Refs fix(ai-slop): let every ignore marker carry the reason the fix flow asks for (0.2.2) #3049 — ai-slop 0.2.2, whose reasoned block markers were the rejected alternative here.

---

Generated by Claude Code

[claude]

Claude finished @kyle-sexton's task in 1m 42s —— View job

---

CI code review — /review:code-review

• Gather context (PR diff, .claude/ai-slop.json history, ai-slop setup contract)
• Understand the request (invoke /review:code-review per the workflow header)
• Review the diff against correctness/maintainability criteria
• Post findings

Skip gate: this change is trivial with no meaningful review surface — a single JSON string-array append plus its accompanying _comment rationale, mechanically identical in shape to the two entries it sits beside. No inline findings.

What I checked (verified myself, not just author-claimed):

• git show 4a268ca -- .claude/ai-slop.json — the diff is exactly "rule-emoji-formatting" appended to disabled_rules and the _comment extended to cover it; nothing else moved.
• rule-emoji-formatting is a real, shipped rule ID — confirmed against plugins/ai-slop/skills/audit/scripts/detect.sh:59 and plugins/ai-slop/skills/audit/reference/catalog.md:262.
• The file still carries exactly the three keys the setup contract recognizes (_comment, excluded_paths, disabled_rules) per plugins/ai-slop/skills/setup/SKILL.md:19-25 — no unknown-key drift.
• The PR's claim that per-rule path allowlisting is supported only for rule-em-dash today is accurate: plugins/ai-slop/skills/setup/SKILL.md:20 documents em_dash_allowed_paths as the only such key; there's no emoji- or curly-artifact-specific equivalent in the setup contract.

Not independently verified (this lane's Bash access is scoped to gh pr * only, so I could not execute the scripts myself): detect.sh --show-config output, the scoped rule-emoji-formatting findings=0 run, detect.test.sh's 92 passing cases, check-changelog-parity.sh, editorconfig-checker, and typos. These are author-claimed in the PR description and should be taken as such rather than as independently confirmed here.

No correctness or maintainability issues found in the reviewed diff.

[claude]

Claude finished @kyle-sexton's task in 44s —— View job

---

Security review

• Gather context (PR diff, changed files)
• Apply skip gate / assess security-relevant surface
• Report findings

Skip gate hit: change has no security-relevant surface.

The diff touches a single file, .claude/ai-slop.json, and only:

1. Adds "rule-emoji-formatting" to the disabled_rules array.
2. Extends the _comment documentation string with the rationale and measurement for that addition.

No code execution paths, no authorization/access-control logic, no secrets/token handling, and no GitHub Actions workflow or permissions changes are touched. This is a data-only config file consumed by the ai-slop detector to decide which lint rules to skip — it carries no injection surface and doesn't alter any trust boundary. Per the skip gate in /review:security-review, no further review is needed and no findings are reported.

[github-actions]

Last security-reviewed head: 4a268cabba37bfc2ab18e13e279717e7016642ce. On the next push, the relevance gate compares only the commits since this SHA; delete this comment to force a full re-review.

[github-actions]

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