Discover testing security llm ai skills, compare source repositories and read the instructions for each workflow.
Assembles coder and verifier prompts from templates. Use when user says "build a prompt", "assemble a prompt", "assemble prompts", or "generate a prompt" for dispatching work to coding and verification agents.
Remove AI writing patterns from text. Based on Wikipedia's AI Cleanup research with 24 pattern categories including inflated symbolism, promotional language, AI vocabulary, and formatting issues.
Local search/indexing CLI (BM25 + vectors + rerank) with MCP mode.
Guide for creating production-ready Agent Skills. Use when creating new skills, improving existing skills, or diagnosing why a skill isn't working effectively. Covers skill quality standards, 5 skill type patte...
Implement DB-persisted conversation state for stateless AI agent servers. This skill should be used when users need to create/load conversations, append messages with role ordering, fetch history for agents, or...
Build AI agents using OpenAI Agents SDK with MCP server integration. Supports Gemini directly via AsyncOpenAI or OpenRouter for non-OpenAI models. Covers agent creation, function tools, handoffs, MCP server con...
Routes to appropriate Ralph skills based on agent role and task signals
Reviews plot structure, story beats, logic flow, and narrative continuity to identify contradictions, plot holes, pacing issues, or unrealistic elements.
Build custom chat API backends for OpenAI ChatKit frontend using FastAPI. Covers chat endpoint implementation, SSE streaming, conversation persistence, and integration with OpenAI Agents SDK + Gemini via LiteLL...
Persistent Python REPL server that avoids reloading heavy packages (torch, transformers, nnterp, vllm). Use when running multiple Python commands with heavy imports, loading ML models, or doing iterative testin...
Complete guide to subagents for project-scoped workers in Claude Code. Use when creating .claude/agents/*.md files or adding context: fork to skills. Do not use for standalone plugin agent development.
Evaluate Agent Skills against official specifications and 8-dimensional scoring framework. Use when: (1) Reviewing SKILL.md files for quality, (2) Auditing skill packages for compliance, (3) Improving existing...
This skill should be used when the user asks to "validate skills", "check skill format", "verify agents", or "audit .claude directory".
Local speech-to-text with the Whisper CLI (no API key).
Scans the .claude/ directory and updates CLAUDE.md to reflect current skills, rules, and agents
Implement single-endpoint chat API with deterministic request orchestration. This skill should be used when users need to create /api/{user_id}/chat endpoint, implement request validation, resolve conversation...
Use CodexBar CLI local cost usage to summarize per-model usage for Codex or Claude, including the current (most recent) model or a full model breakdown. Trigger when asked for model-level usage/cost data from c...
Create self-sufficient skills following Agent Skills standard. Use when building autonomous capabilities with progressive disclosure. Do not use for general programming or non-Claude contexts.
Master coordinator agent that executes tasks by delegating to specialized agents in parallel, validating results, and ensuring quality
Transcribe audio via OpenAI Audio Transcriptions API (Whisper).
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or...
Project-scoped context: fork router for .claude/agents/ with isolation and parallelism patterns. Use for creating, auditing, or refining subagents in current project. Routes to subagents-knowledge for coordinat...
Retrospective contribution format for Developer, Tech Artist, QA, and Game Designer worker agents
Project-scoped skills router for .claude/skills/ with progressive disclosure. Use when creating, auditing, or refining skills in current project. Routes to skills-knowledge for implementation details. Do not us...
The default order uses GitHub repository stars from our indexed data. Stars belong to the source repository and may be shared by multiple skills. They indicate repository interest, not skill installs or an independent quality score.