Orchestrate a full-project consistency audit using a parallel agent team. Spawns lane-specific auditors to cross-reference docs vs code vs tests vs infra, deduplicates findings, and writes a severity-scored audit report.
Run a security audit and dependency check on project dependencies. Use when you want to check for vulnerabilities, audit packages, or update outdated dependencies.
Run a comprehensive codebase hygiene audit. Spawns a team of recon and research agents to analyze structure, dependencies, inconsistencies, and industry best practices, then synthesizes findings into an actionable refactoring plan with phased implementation.
Review a pull request for correctness, convention compliance, and test coverage. Use when you want to review a PR, check code quality, or get feedback on changes before merging.
API development conventions and patterns. Applies when writing or modifying API routes, MCP tools, or CLI commands -- covers authentication, ownership enforcement, error responses, and route organization.
Update architecture documentation after code changes to keep docs in sync with reality. Use when structural changes have been made and documentation needs updating.
Fix a GitHub issue end-to-end -- read the issue, implement the fix, write tests, and create a commit. Use when you want to fix a specific GitHub issue, resolve a bug report, or implement a feature request.
Check the current deployment status, health, and recent CI runs. Use when you want to know if the service is up, check recent CI/CD activity, or diagnose deployment issues.
Memory consolidation skill for Claude Code. Scans session transcripts for corrections, decisions, preferences, and patterns, then merges findings into persistent memory files. Auto-triggers via native Stop hook every 24hrs. Inspired by how sleep consolidates human memory.
Guided ideation with built-in feasibility research. Conversation with the user to explore ideas, research agents validate practicality, then resolve into a requirements doc for /plan-features.
Adversarial 5-agent debate to evaluate what you have. Critics attack, defenders respond, moderator writes verdict. Produces a readiness report -- feed into /plan-features to act on findings.
Execute an execution plan document by orchestrating worktree agents, enforcing commit/merge/test gates, and tracking progress. Main agent stays lean -- all work delegated.
Structured debate -> plan -> execute pipeline for evaluating feature readiness. Runs 5-agent debate, creates action plan with review, then executes fixes.