refactor(skill): tighten request-audit, merge Sujith/Burrasec templates

[0xDEnYO]

Summary

Sub-PR onto #1761. Tightens the request-audit skill from 433 → 356 lines (~18%) without changing behavior. Audit of the diff in request-audit-skill-optimization/ (local-only) confirms every functional bit of the source is preserved.

What was consolidated

• Merge 3b + 3c into one parameterized template. The Sujith and Burrasec thread-reply templates were near-identical — only difference was greeting position. Now one template with a 2-row variant table at the top of the Channels section spells out the per-channel diff (mention ID, display name, greeting position).
• Flatten Step 2 sub-headers (Scope / Context / Urgency) to inline bold labels. Same four scope sources, same four context-extract sources, same three urgency triggers — just less heading noise.
• Collapse Setup section while preserving the .env example, the WEBHOOK_<CHANNEL> convention, and the 1Password reference verbatim.
• Worked PR feat: migrate GenericSwapFacetV3 to granular allowlist + LibAsset routing [GenericSwapFacetV3 v2.0.0] #1715 example: keep the full Tron context paragraph (that's the prose-density calibration — cutting it would lose functionality). Burrasec block collapsed to a single sentence since the variant table now defines the diff.

What was not removed

All preserved verbatim:

• Channels table with both Slack IDs (C08CJ4FEHQA / C094C46JSDP) + the "re-verify" warning
• Step 1 fetch — all 8 --json fields, the "last entry's oid" rule, the empty-array re-fetch
• Step 1b — Slack URL parsing + four synthesis questions + free-form + skip branches
• Code-style rule (full list of token types + examples)
• Step 4 preview format — quad-backtick outer fence + triple-backtick inner fences, exact same rendered layout
• Steps 5 + 6 exit-code semantics + manual-fallback transforms (display-name + plain-text + single-write rule + format) + the inconsistent 1Password wording from both source locations (kept both as-is — not in scope to fix)
• Error-handling table

Pre-flight

• coderabbit review --base origin/feat/request-audit-skill --plain → No findings ✔ on first pass
• No tests / build affected (markdown-only change to .agents/commands/request-audit.md; .claude/skills/... and .cursor/commands/... are symlinks to the canonical file)

Test plan

• Sanity-read the rendered skill on this branch — does the merged thread-reply template still produce the same two outputs the source produced?
• (Optional) Run /request-audit <some-PR> against this branch and compare the drafted parent + thread reply against the original
• If something feels too compressed, ping me — I can selectively restore individual sections

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: ddb2a0ac-b6eb-45ab-b5ff-f68a031a47fe

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch chore/request-audit-skill-tighten

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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