review-automation-patterns
Installation
SKILL.md
Review Automation Patterns
Overview
Machines enforce rules with no exceptions (formatting, unused variables, known CVEs, type mismatches). Humans evaluate intent, architecture fitness, naming coherence, test quality, and logic correctness. The goal of review automation is to eliminate machine-checkable items from human review.
Pairs with review-efficiency-patterns and review-code-quality-process. Cross-reference security-patterns-code-review for security-specific automated checks.
Tool Selection by Language
| Language | Linter | Type Checker | Security Scanner | Formatter |
|---|---|---|---|---|
| Go | golangci-lint, go vet | go build (compile-time) | gosec | gofmt, goimports |
| Python | ruff (replaces flake8+isort+pyupgrade) | mypy, pyright | bandit, safety | black, ruff format |
| TypeScript | eslint + typescript-eslint | tsc --noEmit | semgrep, npm audit | prettier |
| Rust | clippy | rust-analyzer, rustc | cargo-audit, cargo-deny | rustfmt |
| Java | checkstyle, spotbugs, PMD | javac (compile-time) | snyk, OWASP dependency-check | google-java-format, spotless |
| C++ | clang-tidy, cppcheck | clang (compile-time) | — (use manual review) | clang-format |