How do I install the Review Fix skill?

Run npx skillmds@latest add nwiizo/review-fix-2 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Review Fix safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Review Fix?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Review Fix free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Review Fix?

nwiizo (@nwiizo) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Review Fix by nwiizo — Agent Skill

Name: Review Fix
Author: nwiizo

File contents

Review uncommitted changes, fix actionable issues, and verify the result.

Workflow

Phase 1: Scope

git status --short and git diff --stat HEAD to determine what changed.
If no changes exist, abort with "nothing to review".

Phase 2: Baseline

Run build and lint to establish a clean starting point:

cargo clippy --workspace -- -D warnings 2>&1 | tail -20
cargo test --workspace --no-run 2>&1 | tail -5

If the baseline already fails, report the failures and stop — do not mask pre-existing breakage.

Phase 3: Review

Inspect the diff and new untracked files. Check for:

Category	What to look for
Stale references	Deleted modules/commands still referenced in routing, docs, templates
Code duplication	Identical structs/functions across multiple files (> 10 lines, > 90% similarity)
Provider abstraction gaps	Hard-coded CLI names (`claude`, `codex`, `gh copilot`) instead of `AgentProvider` trait
Dead async	Async operations whose results are discarded (`let _ = tokio::...await.ok()` then re-done synchronously)
Sort instability	`Utc::now()` used as fallback sort key (non-deterministic)
Hardcoded paths	Filesystem paths that should be relative to `repo_path`
Unused TODO params	CLI arguments accepted but silently ignored (`_param: T`)
Security	Path traversal in user-supplied IDs, credential leaks

Classify each finding:

🔴 Fix: Clear bug or correctness issue — fix it now.
🟡 Warn: Design concern or missing feature — report but do not fix unless trivial.
🟢 Note: Style or informational — report only.

Phase 4: Fix

For each 🔴 finding:

Make the minimal, surgical edit.
Do not change unrelated code.
After all fixes, run:

cargo clippy --workspace -- -D warnings
cargo test --workspace --no-run

If clippy or test compilation fails, revert the offending edit and report it as 🟡 instead.

Phase 5: Report

Output a structured summary:

## Review + Fix Report

### Fixed (🔴)
- <file>:<line> — <description>

### Warnings (🟡)
- <file>:<line> — <description>

### Notes (🟢)
- <description>

### Verification
- clippy: PASS / FAIL
- test compile: PASS / FAIL
- files modified: N

---name: review-fix-2description: Code review with automatic fix. Detects bugs, stale references, code duplication, provider abstraction gaps, and sort instability — then fixes and verifies.---Review uncommitted changes, fix actionable issues, and verify the result.## Workflow### Phase 1: Scope1. `git status --short` and `git diff --stat HEAD` to determine what changed.2. If no changes exist, abort with "nothing to review".### Phase 2: BaselineRun build and lint to establish a clean starting point:```bashcargo clippy --workspace -- -D warnings 2>&1 | tail -20cargo test --workspace --no-run 2>&1 | tail -5```If the baseline already fails, report the failures and stop — do not mask pre-existing breakage.### Phase 3: ReviewInspect the diff and new untracked files. Check for:| Category | What to look for ||----------|-----------------|| **Stale references** | Deleted modules/commands still referenced in routing, docs, templates || **Code duplication** | Identical structs/functions across multiple files (> 10 lines, > 90% similarity) || **Provider abstraction gaps** | Hard-coded CLI names (`claude`, `codex`, `gh copilot`) instead of `AgentProvider` trait || **Dead async** | Async operations whose results are discarded (`let _ = tokio::...await.ok()` then re-done synchronously) || **Sort instability** | `Utc::now()` used as fallback sort key (non-deterministic) || **Hardcoded paths** | Filesystem paths that should be relative to `repo_path` || **Unused TODO params** | CLI arguments accepted but silently ignored (`_param: T`) || **Security** | Path traversal in user-supplied IDs, credential leaks |Classify each finding:- 🔴 **Fix**: Clear bug or correctness issue — fix it now.- 🟡 **Warn**: Design concern or missing feature — report but do not fix unless trivial.- 🟢 **Note**: Style or informational — report only.### Phase 4: FixFor each 🔴 finding:1. Make the minimal, surgical edit.2. Do **not** change unrelated code.3. After all fixes, run:```bashcargo clippy --workspace -- -D warningscargo test --workspace --no-run```If clippy or test compilation fails, revert the offending edit and report it as 🟡 instead.### Phase 5: ReportOutput a structured summary:```## Review + Fix Report### Fixed (🔴)- <file>:<line> — <description>### Warnings (🟡)- <file>:<line> — <description>### Notes (🟢)- <description>### Verification- clippy: PASS / FAIL- test compile: PASS / FAIL- files modified: N```

Frequently asked questions

Review Fix