review-crate — Claude Code skill: complete Rust crate audit
review-crate reads every file in a Rust crate, runs tests, and reports bugs with exact line numbers. Use it to catch logic errors, panics, race conditions, and validation gaps that unit tests miss.
no license 115 installs ★ 5 updated 2026-08-31
npx skills add ulpi-io/skills --skill review-crate -a claude-codeWhat review-crate does
review-crate enumerates and reads every file—from Cargo.toml through source, tests, and local docs. It then runs cargo test and analyzes the results. Each finding includes category (correctness, panic, validation, race, data-corruption), severity, and the exact code location: file:line. Before writing to the issue file .ulpi/issues/<crate-name>.md, it checks for duplicates.
The skill runs in its own forked analysis thread with a separate token budget, so it does not starve the main workflow. Every claim must ground in real code; uncertain conclusions are marked INFERENCE with an explanation. It does not invent bugs from intuition or write findings outside your repository.
Use it before a release or when you suspect hidden problems in a single crate. It is not a multi-crate sweep tool and does not cover ad-hoc diff review.
What it helps with
- identify logic bugs and data validation failures in a Rust crate
- spot race conditions and state management issues
- find code paths that can panic or crash
- detect gaps in test coverage for risky code
- document issues with exact file and line references
Example prompts for the agent
- /review-crate crates/hgdb-storage
- Deep audit this crate—read all files, run tests, report findings
- Review the entire async-pool crate end to end and write the issue file
How to install review-crate
1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill review-crate -a claude-code2Manually: put the files in place
~/.claude/skills/review-crate/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/review-crate/
1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill review-crate -a codex2Manually: put the files in place
.agents/skills/review-crate/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill review-crate -a cursor2Manually: put the files in place
.cursor/skills/review-crate/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ulpi-io/skills review-crate2Manually: put the files in place
.github/skills/review-crate/SKILL.md1Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill review-crate -a gemini-cli2Platform CLI
gemini skills install https://github.com/ulpi-io/skills --path review-crateManually: put the files in place
.devin/skills/review-crate/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ulpi-io/skills --skill review-crate -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about review-crate
How do I install review-crate in Claude Code?
Run npx skills add ulpi-io/skills --skill review-crate -a claude-code, then open ~/.claude/skills/review-crate/SKILL.md and copy its content into Claude Code.
What does review-crate do?
It reads every file in a crate, runs tests, analyzes the code for bugs (panics, race conditions, validation errors, data corruption), and writes findings to a local issue file with exact line numbers.
Why not just run cargo test?
cargo test catches test failures, but it does not analyze your source code for logic bugs, race conditions, or validation gaps. review-crate reads the entire codebase, runs tests, and reports real issues with precise locations.