Skip to content

modrexio/modrex / .agents/skills/ai-review

ai-review

Verified

Review a diff for AI-shaped code before a PR, commit, or final summary

ai-review

Use this skill before proposing a PR, commit, or final response for a non-trivial code change.

Start with blocker-level issues from .agents/rules/ai-danger-patterns.md.

Review target

Default to git diff HEAD.

If the user provides a path or range, review only that target.

Blocker checklist

Check first for the five dangerous AI patterns:

  1. Silent catch / silent fallback
  2. Generic helpers replacing game-specific logic
  3. Defensive checks that hide broken invariants
  4. Deep if/else nesting
  5. Fake abstractions / speculative future-proofing

Cleanup checklist

Then check for:

  • obvious WHAT-comments
  • useful structural/header comments wrongly removed
  • any casts
  • unnecessary useMemo or useCallback
  • derived data stored in React state
  • unrelated formatting
  • hardcoded renderer colors (use semantic tokens from index.css)
  • direct API calls that bypass project caches (modCache.ts, thumbnailCache.ts)
  • bare JSX strings (use t('key'))
  • missing Tauri command wiring (all three places)
  • user-facing behavior missing from CHANGELOG when appropriate
  • code that contradicts the nearest AGENTS.md above it

Severity

Use three levels:

  • Blocker: dangerous AI pattern, likely bug, broken invariant, wrong project pattern, unsafe filesystem/archive behavior.
  • Cleanup: maintainability issue, obvious AI-shaped code, useless comment, unnecessary abstraction.
  • Note: acceptable but worth mentioning.

Output

Return:

  1. Blockers
  2. Cleanups
  3. Notes
  4. Suggested next command or skill

If there are no issues:

Diff is clean against the ai-review checklist.

Do not invent problems. Cite the exact file and function when possible.

receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 321kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 321kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 321kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 321kUpdated Claude CodeCodexVerifiedInstall