aka. Agent Skills
Discover skills for AI coding agents. Works with Claude Code, OpenAI Codex, Gemini CLI, Cursor, and more.
Validate documentation claims against deployment reality. Use when checking if docs accurately describe what's deployed, finding false claims, or auditing doc accuracy. Triggers: "verify docs", "doc audit", "check doc claims", "docs match reality", "validate doc status".
Use when: "Python", "Go", "Golang", "Rust", "Java", "TypeScript", "shell script", "bash", "async/await", "goroutines", "ownership", "virtual threads", "type system", "decorators", "generators", "channels", "interfaces", "Spring Boot", "POSIX".
Execute safe, reproducible Git workflows with MANDATORY semantic commits (feat/fix/docs), auto-injected templates, and git hooks integration
Deep codebase exploration. Triggers: "research", "explore", "investigate", "understand", "deep dive", "current state".
Use when: "incident", "outage", "production issue", "debug", "logs", "error", "postmortem", "root cause", "triage", "escalation", "P0", "P1", "on-call", "mitigation", "rollback".
Extract learnings from completed work. Trigger phrases: "run a retrospective", "extract learnings", "what did we learn", "capture lessons", "create a retro".
Quick status check of current work state. Use when the user asks "what's my status", "show status", "where am I", "current state", or needs a snapshot of active work, ready issues, and git state.
This skill should be used when the user asks to "validate changes", "run validation gate", "check before merge", "quality check", or needs comprehensive validation across security, quality, and architecture.
Use when: "security", "penetration test", "pentest", "vulnerability", "OWASP", "network", "firewall", "SSL", "TLS", "DNS", "load balancer", "CDN", "encryption", "authentication", "authorization", "secrets".
Use when: "API", "backend", "frontend", "fullstack", "mobile", "iOS", "React Native", "Flutter", "deploy", "CI/CD", "Docker", "Kubernetes", "GitHub Actions", "LLM", "RAG", "prompt engineering", "microservices", "database schema", "SwiftUI".
This skill should be used when the user asks to "build container", "create Dockerfile", "configure podman", or needs guidance on OpenShift container images, arbitrary UIDs, or image tagging.
Deep project analysis for architecture planning. Use when starting migration or designing new agent components for an unfamiliar codebase.
Agent communication patterns. Send mail, check inbox, broadcast to workers, nudge stuck polecats, and escalate to humans when needed.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Use when: "accessibility", "WCAG", "a11y", "customer support", "FAQ", "tickets", "UI", "UX", "design system", "wireframe", "Obsidian", "knowledge graph", "links", "task decomposition", "workflow", "risk assessment", "edge cases", "failure modes".
Scaffold and audit OSS documentation for open source projects. Triggers: "add OSS docs", "create README", "setup contributing", "add changelog", "missing SECURITY.md", "oss documentation", "prepare for open source", "github templates", "add AGENTS.md".
Understanding Gas Town role responsibilities. Clarifies when Claude should behave as Mayor, Crew, Polecat, Witness, or Refinery.
This skill should be used when the user asks to "create documentation", "write a corpus document", "create a standard", "add corpus content", "write GraphRAG documentation", or mentions corpus/standards documentation.
Gold standard for code-map feature documentation. Use when generating docs in docs/code-map/ or running /doc on CODING repositories.
Analyzes and refactors all skills in the project - consolidate, optimize, and remove unnecessary ones. Use when skills need cleanup or reorganization.