# Rust Review > Review Rust changes for correctness, ownership, API ergonomics, error handling, unsafe boundaries, async behavior, tests, performance, features, dependency choices, and greenfield conventions. Use for review-only Rust tasks. Use when this capability is needed. - Skill: `tomevault-io/rust-review-4` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/rust-review-4` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/rust-review-4/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Integrations & APIs - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/rust-review-4 --- # Rust Review ## Rule Provide actionable, evidence-backed findings prioritized by correctness, safety, public contracts, and maintainability. Do not rewrite code unless explicitly asked. ## Review Focus Check ownership design, lifetime complexity, public API ergonomics, error types, panic paths, `unsafe`/FFI justification, feature flags, dependency choices, async cancellation, task lifecycle, tests, docs, security boundaries, observability fields, and performance risks. ## Antipatterns To Flag - `unwrap`/`expect` on user input, I/O, network, database, or request path failures. - Public APIs overfit to internal lifetimes or expose implementation details. - Traits with a single implementation and no clear boundary value. - Default features that surprise library users. - Tokio runtime creation inside libraries. - `Arc>` used as a design shortcut. - Unsafe blocks without local safety comments and tests/Miri evidence where practical. ## Workflow 1. Inspect the diff, `Cargo.toml`, features, tests, and nearby code. 2. Run or recommend focused checks: `cargo test`, `cargo clippy`, `cargo fmt --check`, `cargo deny check`, `cargo audit`, or `cargo nextest` when configured. 3. Identify blocking issues before optional style suggestions. 4. Tie findings to concrete code and explain impact. 5. Suggest small idiomatic fixes. ## Completion Return findings by severity, validation performed or skipped, and areas not reviewed. --- > Source: [nyquistwilder/personal-pi](https://github.com/nyquistwilder/personal-pi) — distributed by [TomeVault](https://tomevault.io).