SuperAPI AI Directory

duck-review

Coding & development

View repository

Updated 10-16 06:39

About this skill

Review 🦆. Extends caveman-review. Keep terse format by default.

Workflow

  1. Confirm review input exists (diff, PR text, or pasted code chunk).
  2. Scan in priority order: security → correctness → data integrity → performance → tests → docs.
  3. Emit only actionable findings. One line each: location, problem, fix direction.
  4. Use strongest matching prefix. If multiple apply, pick highest risk prefix.
  5. For security or irreversible-risk findings, switch to full paragraph (Auto-Clarity), then resume terse comments.

If prefix choice unclear or reviewer needs wording examples, load references/review-comment-examples.md.

Output Format

One-line comment template:

<prefix> <path[:line]> — <problem>. Fix: <smallest safe change>.

Keep comments paste-ready for PR threads.

Prefixes

All caveman-review prefixes apply. Add these:

  • 📝 doc: — missing/outdated docs or annotations
  • 🧪 test: — missing/outdated test coverage
  • 🔒 sec: — security issue (injection, auth bypass, secrets, SSRF)
  • ⚡ perf: — performance concern (N+1, unnecessary alloc, bad complexity)

Input → Output Examples

Input: "review this diff"

Output:

  • 🧪 test: src/auth/session.ts:88 — refresh-token expiry path untested. Fix: add test for expired refresh token returning 401.
  • 📝 doc: api/openapi.yaml:210 — response schema omits new error_code field. Fix: document field in 401 schema.

Input: "review this" + SQL string concatenation with user input

Output:

  • 🔒 sec: db/userRepo.ts:44 — SQL built from raw user input enables injection. Fix: parameterize query placeholders and bind values.

Edge Cases

  • No diff or code provided: ask for concrete review target before commenting.
  • Huge refactor: report highest-impact findings first; avoid line-noise nits.
  • Uncertain finding: ask one clarifying question instead of inventing certainty.
  • Same line has multiple problems: split into separate comments when fixes differ.
  • Security finding mixed with style issues: report security first, style optional.

Auto-Clarity

Drop terse mode for security findings, architectural disagreements, onboarding contexts. Write full paragraph there, resume terse after.

Boundaries

Reviews only. Don't write patch, don't approve/request-changes, don't run linters/tests. Output comments ready to paste into PR.