LLM & Agents

Install WebMind MCP for DOM analysis. Use when user needs Ragic DOM assistant capabilities.

Keep project documentation accurate when behavior changes. EXCLUSIVE to project-manager agent.

Run an iterative, hypothesis-driven “debug mode” loop to diagnose and fix a bug using targeted instrumentation and log capture. Use when the user wants an interactive debug loop, when you need to quickly narrow a failure via added debug statements, or when you need a lightweight way to centralize logs from a repro run (via `agent-skills debug` server + SSE UI).

Configure and use Supabase MCP from Claude Code to inspect and manage a Supabase project. Use when the user asks to create tables, migrations, RLS policies, storage buckets, or to query Supabase from Claude Code. Includes safe setup steps (project-scoped, dev-only, optional read-only).

Use when answering complex questions about a codebase that require exploring multiple areas or understanding how components connect - coordinates parallel sub-agents to locate, analyze, and synthesize findings

MagentaA11y accessibility acceptance criteria reference. Use this skill when users ask for accessibility criteria, acceptance criteria (AC), Gherkin tests, developer notes, or how to test components for accessibility. Supports both Web and Native (iOS/Android) platforms. Triggers on requests like "give me the Button AC", "how to test a checkbox", "accessibility criteria for modal", "native button Gherkin", "web link developer notes", or any MagentaA11y component query.

新しいスキルを対話形式で作成するメタスキル。 スキルの作成、新規スキル追加、カスタムスキル開発、スキルテンプレート生成をサポート。 要件をヒアリングし、SKILL.md、scripts/、references/を自動生成する。 Use when user asks to create a new skill, add custom skill, or develop agent capabilities.

Permission validation, architecture patterns, anti-patterns, and best practices for Claude Code settings management

Replace with description of the skill and when Claude should use it.

Guide for creating custom agents for Claude Code with specialized behaviors and tools

End-to-end system for creating supervised fine-tuning datasets from books and training style-transfer models. Covers text extraction, intelligent segmentation, synthetic instruction generation, Tinker-compatible output, LoRA training, and validation.

Programmatic browser automation using Playwright MCP server. Use when building Claude Code tools or applications that need web automation, testing, scraping, or browser interaction. Provides structured accessibility-based automation without screenshots or vision models.

Use this skill when building AI voice agents with the ElevenLabs Agents Platform. This skill covers the complete platform including agent configuration (system prompts, turn-taking, workflows), voice & language features (multi-voice, pronunciation, speed control), knowledge base (RAG), tools (client/server/MCP/system), SDKs (React, JavaScript, React Native, Swift, Widget), Scribe (real-time STT), WebRTC/WebSocket connections, testing & evaluation, analytics, privacy/compliance (GDPR/HIPAA/SOC 2), cost optimization, CLI workflows ("agents as code"), and DevOps integration. Prevents 17+ common errors including package deprecation, Android audio cutoff, CSP violations, missing dynamic variables, case-sensitive tool names, webhook authentication failures, and WebRTC configuration issues. Provides production-tested templates for React, Next.js, React Native, Swift, and Cloudflare Workers. Token savings: ~73% (22k → 6k tokens). Production tested. Keywords: ElevenLabs Agents, ElevenLabs voice agents, AI voice agents, conversational AI, @elevenlabs/react, @elevenlabs/client, @elevenlabs/react-native, @elevenlabs/elevenlabs-js, @elevenlabs/agents-cli, elevenlabs SDK, voice AI, TTS, text-to-speech, ASR, speech recognition, turn-taking model, WebRTC voice, WebSocket voice, ElevenLabs conversation, agent system prompt, agent tools, agent knowledge base, RAG voice agents, multi-voice agents, pronunciation dictionary, voice speed control, elevenlabs scribe, @11labs deprecated, Android audio cutoff, CSP violation elevenlabs, dynamic variables elevenlabs, case-sensitive tool names, webhook authentication

Quality gate agent that verifies plans against project history and architectural constraints.

Manage Obsidian vault for persistent knowledge storage, search, and retrieval

Master Claude Code plugin architecture, design patterns, and technical implementation. Learn to build scalable, maintainable plugins with proper structure, performance optimization, and security best practices.

Error handling patterns and recovery protocols for agents

Use Claude Code's interactive question tool to gather comprehensive requirements through structured multi-select questions

MCP server management for Claude Desktop. Profile-aware - reads MCP server inventory from profile.mcp.servers{} and config path from profile.paths.claudeConfig. Use when: installing MCP servers, configuring Claude Desktop, troubleshooting MCP issues.

Add additional test cases to an existing exercise to improve coverage