LLM 與 Agent

Multi-agent and MCP pipeline security with 5-layer defense architecture. Use when building MCP servers, multi-agent systems, or any pipeline that handles user input to prevent prompt injection and ensure proper authorization.

Why pinning GitHub Actions to SHA-256 commits matters for supply chain security. Attack vectors from unpinned actions and comparison of tag vs SHA pinning.

Guide for building production-ready LiveKit voice AI agents with multi-agent workflows and intelligent handoffs. Use when creating real-time voice agents that need to transfer control between specialized agents, implement supervisor escalation, or build complex conversational systems.

A skill for creating comprehensive Python unit tests using pytest. It provides guidance and templates for test structure, fixtures, mocking, parameterization, and coverage. Use this when you need to write unit tests for Python code, set up a testing environment with pytest, or apply testing best practices.

Create hierarchical project plans optimized for solo agentic development. Use when planning projects, phases, or tasks that Claude will execute. Produces Claude-executable plans with verification criteria, not enterprise documentation. Handles briefs, roadmaps, phase plans, and context handoffs.

Run repo startup: prompt load, checklist sync, optional PR triage, worktree hygiene, and CI pull. Best-effort with structured status.

Apply approved remediation actions from audit resolution. Implements config file updates, template syncs, and code fixes using appropriate domain/config agents. Use when audit findings have been approved for remediation and need implementation.

Analyzes Claude Code session history to identify repeated workflows and suggest slash commands to automate them

Build AI gateway services for routing and managing LLM requests. Use when implementing API proxies, rate limiting, or multi-provider AI services.

Standards for Claude-optimized repository documentation. Use when creating or updating docs/claude/ documentation to ensure correct format, abstraction level, and front-matter.

Create Claude Code subagents. Use when user wants to create a subagent, specialized agent, or custom AI assistant for Claude Code.

Performs automatic web search using Gemini CLI with Google Search grounding. Triggers on "Gemini search:", "suche mit Gemini", "use Gemini to find". Claude writes a request file, the Watcher executes the search, Claude reads the result. FREE Gemini 2.5 Pro with 1000 requests/day.

Use when spec is complete and you need detailed implementation tasks for LLM agents. Creates execution plans with exact file paths, complete code examples, and verification steps. Triggers: 'write plan', 'execution plan', 'implementation plan', 'break down into tasks', 'detailed steps'.

Guide to creating agent skills with proper structure and best practices

Use when starting a fine-tuning project to determine if fine-tuning is needed, or when evaluating whether a base model meets quality thresholds for a specific domain task

Examples and patterns for writing effective agent and skill descriptions. Use when crafting descriptions that serve as routing keys for Claude's invocation decisions.

Search RAG database for relevant content. Use for semantic queries over processed documents, code, or papers.

Agent Lightning를 사용하여 AI 에이전트를 자동으로 최적화하는 방법을 제공합니다.

Guide Claude in writing efficient, idiomatic SWI-Prolog DCGs (Definite Clause Grammars) following best practices for single-pass parsing, character codes, pure declarative style, and accumulator patterns. Use when working with Prolog parsing tasks.

Generate and run tests for TypeScript/JavaScript code using Bun test runner. Use this skill when you need to write tests, check coverage, or debug test failures.