Agent Claude Code
Use this agent as the second stage of a code review, after spec compliance is confirmed — evaluating code quality, security vulnerabilities, performance issues, and test coverage without re-checking requirements.
Agent Claude Code
Use this agent to verify that a pull request or set of changes fully implements its stated requirements — checking requirements coverage, missing features, and scope creep — before evaluating code quality.
Agent Claude Code
Use this agent to perform a thorough two-stage review of a pull request or set of changed files — first verifying spec compliance, then evaluating code quality, security, test coverage, and performance. Prefer this over the inline /review command when the diff spans more than 5 files or more than 300 lines.
Agent Claude Code
Use this agent to audit all dependencies across a repository for vulnerabilities, outdated packages, unpinned versions, and abandoned libraries. Invoke when the user asks to "audit dependencies", "check for vulnerable packages", "update dependencies", or "scan for outdated libraries". Prefer this over a manual npm…
Agent Claude Code
Use this agent to generate a comprehensive new-joiner guide for a repository — covering what the service does, how to run it locally, architecture overview, key files, environment setup, gotchas, and who owns what. Invoke when the user asks to "write an onboarding guide", "create a new-joiner doc", "document this repo…
Agent Claude Code
Use this agent to perform a full codebase security audit covering OWASP Top 10 vulnerabilities — hardcoded secrets, injection flaws, broken auth, weak crypto, misconfiguration, and vulnerable dependencies. Invoke when the user asks to "audit", "scan", or "security review" a directory, service, or codebase, especially…
Agent Claude Code
Use this agent to analyse an entire module or directory for missing test coverage, then generate the missing tests. Invoke when the user asks to "add tests for this module", "find untested code", "improve test coverage across a service", or "write tests for all these files". Prefer this over the inline /test-gen…
Command Claude Code
Generate an Architecture Decision Record (ADR) from a discussion or decision.
Command Claude Code
Generate a changelog from git history since the last release tag.
Command Claude Code
Create one or more GitHub issues from a description. Accepts natural language — infers titles, labels, and bodies from context.
Command Claude Code
Run a pre-deployment verification checklist before releasing to production.
Command Claude Code
Explain a code file, function, or architectural pattern in detail with context.
Command Claude Code
Save a curated session-level summary to history. Use this at meaningful milestones — end of session, before a long break, after wrapping a chunk of work — not after every turn. The auto-hooks (UserPromptSubmit, Stop, PreCompact) already capture turns; this command is for the higher-signal arc that those per-turn saves…
Command Claude Code
Generate a postmortem document from an incident description.
Command Claude Code
Generate a Product Requirements Document (PRD) pre-filled from context.
Command Claude Code
Analyze code for smells and suggest or apply refactoring with examples.
Command Claude Code
Fetch a GitHub issue, apply the fix, comment with what was done, and close the issue.
Command Claude Code
Review the current changes (staged, unstaged, or a specific PR) against coding standards and security best practices.
Command Claude Code
Deliver a brutally honest, technically sharp roast of the current repository. Spot real bugs, code smells, security holes, and engineering sloppiness. Be pointed and witty, but every finding must be accurate and include a fix.
Command Claude Code
Scaffold a new service or project with production-ready boilerplate.
Command Claude Code
Audit code against the security skill checklist and OWASP best practices.
Command Claude Code
Run local CI validation on all skills and agents in the spellbook — same checks as .github/workflows/ci.yml.
Command Claude Code
Generate unit and integration tests for a given file or function.
Command Claude Code
Read .claude/findings.jsonl, show pending skill findings, and apply approved changes back to the relevant skill file.