Skills for Real Engineers. Straight from my .agents directory.
codebase-audit
Agent skill fromTencentCloudBase/CloudBase-AI-Toolkit
Perform a full codebase review, categorize findings by severity, file GitHub issues, then fix each issue in an isolated git worktree and submit PRs. Use this skill when the user asks to audit the codebase, do a comprehensive code review, find and fix security/quality/reliability issues, or run a proactive health check across the entire repository.
How to install codebase-audit
npx skills add TencentCloudBase/CloudBase-AI-Toolkit# or add only this skill to Claude Code: copy the folder
skills/codebase-audit/
# into
~/.claude/skills/codebase-audit/ (all projects)
.claude/skills/codebase-audit/ (one project)
1# Codebase Audit → Issue → Worktree Fix → PR23End-to-end workflow: systematically review the entire codebase, report findings as GitHub issues, fix each issue in an isolated git worktree, and submit PRs — all in one session.45## When to use this skill67Use this skill when you need to:89- Perform a full code review / audit of the codebase10- Proactively find security vulnerabilities, logic bugs, or code quality problems11- Turn code review findings into tracked GitHub issues12- Fix each issue in isolation (worktree per issue) and submit PRs13- Run a periodic codebase health check with automated follow-through14- Audit and fix dependency security vulnerabilities (Dependabot alerts / npm audit)1516**Do NOT use for:**1718- Reviewing or fixing a single known bug (use `systematic-debugging` or direct fix)19- Triaging existing open PRs (use `pr-review-fix`)20- Processing attribution issues (use `mcp-attribution-worktree`)21- Feature development or refactoring unrelated to audit findings2223## Workflow2425### Phase 1 — Review26271. Read `references/review-strategy.md` for the review scope and checklist.282. Use the `code-explorer` subagent to read ALL source files in the target directory (default: `mcp/src/`).293. For each file, systematically check against the review checklist:30 - **Security**: path traversal, injection, unvalidated input, hardcoded secrets, improper error exposure31 - **Error handling**: missing try-catch, swallowed errors, error messages leaking internals32…
Excerpt shown under the repository's MIT license.
More skills in CloudBase-AI-Toolkit
Similar skill repositories
An agentic skills framework & software development methodology that works.
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
