Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Creates code-based evaluators for LangSmith-traced agents. Use when building custom evaluation logic, testing tool usage patterns, or scoring agent outputs programmatically. Triggers on requests to evaluate agents, create evaluators, or run experiments against LangSmith datasets.
Teach Claude ANY topic - code libraries, APIs, concepts, tools, methodologies, or domains. Researches via web and docs, then retains knowledge as a permanent skill. Use when user says "/learn <topic>", "learn about X", "teach yourself Y", "become an expert on Z". Examples - "/learn stripe" for payments, "/learn GTD" for productivity, "/learn israeli-tax-law" for domain knowledge.
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 services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
Search, install, configure, update, and remove MCP servers across coding agents (Claude Code, Cursor, VS Code, Claude Desktop, Gemini CLI, Codex, Goose, Zed, Devin CLI/Desktop, and more). Supports multi-agent installation via npx add-mcp, the official MCP registry, and direct config editing.
Generate an /insights-style HTML report across MANY of your agent coding sessions (Claude Code and OpenAI Codex): what you work on, how you use the agent, your wins, your friction points, and concrete fixes to become more effective — precise goal-setting, not trusting output blindly, scoping, context hygiene, right-sizing models. Reads local session history and writes one self-contained report.html. Use when the user asks to "разбери мои сессии", "оцени как я работаю с агентом", "мой insights", "session insights report", "review how I work", "на что обратить внимание", or runs "/mentor", "/mentor claude", "/mentor codex", "/mentor both". Do NOT use for reviewing the CODE produced (that's code-review) or building the agent.
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 codexbar, or when you need a scriptable per-model summary from codexbar cost JSON.
Supplies the approval rule another RHDH skill applies when it is already about to change something outside the session: how to state each operation, what approval binds to, what to report afterwards, and how to keep credentials out of a plan preview. Cited by name from the skill doing the work. Not an entry point — it performs no forge, Jira, or repository action itself, and a request to open, comment, transition, push, or post belongs to the skill that owns that target.
Designs, builds, debugs, and documents OpenClaw workflows, skills, and AI assistant configurations. Use when the user mentions "OpenClaw," "personal AI assistant," "local AI," "ClawdHub," "openclaw skills," "chat platform AI," or wants to set up AI assistants across WhatsApp, Telegram, Discord, or Slack.
Think through ideas, investigate problems, and clarify requirements before committing to a change using `/opsx:explore`. Use when the user says "explore an idea", "think through this", "investigate options", or wants to brainstorm before creating a formal change.
Internal pack contracts for Gabriel Lafrance Skills: asking, stateless execution context, subagent dispatch, review behavior, and how every agent opens a PR. Not user-invoked. Other skills link here so shared docs install with npx skills (skill folders only).
Collaboratively design, implement, review, and refine software features with a human developer using an iterative pair programming workflow. Prioritize understanding, explicit approval, small reviewable changes, and synchronized specifications.
A team skill that designs, implements, and verifies a single feature through a 3-agent (Architect→Builder→Reviewer) loop. Triggered by the keywords "3a로 만들어줘", "3에이전트", "설계-구현-검토", "team-3a". Lighter than peach-team-dev, suited to single features and small-scale changes.
Break down a user story into multiple testable tasks. Use when the user mentions: story, ticket, task breakdown, implement feature, plan implementation, create tasks, split into tasks, story breakdown, task decomposition, acceptance criteria, definition of done, sprint planning, backlog, user story, feature breakdown, ralph, ralph loop, autonomous implementation, .TASK.md, progress.txt, or run through ralph. Each task must be independently verifiable. NOTE: Ralph-loop task files (.TASK.md) are generated ONLY when explicitly requested — never automatically during normal story creation.
Rewrites a rough prompt, voice dump, or half-written instruction into an agent-ready prompt. Use when the user asks to make a prompt better, enhance/refine a prompt, clean up instructions for an agent, or improve pasted prompt text. One-shot transformer; does not interview.
Generates CHANGES.md — an operational index of all OpenSpec changes for a project, with dependency tree, parallelism gates, critical path, multi-agent plan, and per-change scope, governance level, dependencies and "Leer antes" pointers to the knowledge base. Fire-and-forget, no questions asked. Trigger: When user asks to create, build, regenerate, or update CHANGES.md, roadmap, change map, implementation plan; or says "armar CHANGES", "armar roadmap", "crear mapa de changes", "generar plan de implementación", "qué changes necesito", "índice de changes".
Use when a founder wants a hard business decision debated by a council of world-class operators, says "roast room", "put this in the Roast Room", "founder council", "convene the council", "take this to the board", "what would Hormozi, Bezos, or Buffett say", "should I do X", or can't make up their mind on pricing, hiring, raising money, a new product, a partnership, a pivot, an event, or spending. A board of sub-agents, each built from one expert's public work (Bezos, Paul Graham, Hormozi, Gary Vaynerchuk, Dan Martell, Gino Wickman, and Buffett out of the box, plus any seat added with /newroaster), argue it over three rounds, then a Chair returns a scored verdict, the reasoning, the dissent, and next steps. Also does a one-seat consult ("ask the Hormozi seat"). Not /roast (generic personas, idea validation).
💰 Save Token | Token Saver: a guiding skill that helps agents identify and avoid sending duplicate context to LLM APIs. Teaches agents to recognize repeated content and summarize instead of re-sending. Guides agents to recognize duplicated content and avoid re-sending it, thereby saving tokens.
Scaffold a new SciAgent-Skills entry. Picks pipeline/toolkit/database/guide template, creates skills/{category}/{name}/SKILL.md with valid frontmatter, appends the registry.yaml entry, runs validation. Enforces name uniqueness, kebab-case, description keyword rules, schema rules from AGENTS.md. TRIGGER when user says (any language): "add a SciAgent skill", "add a skill for <X>", "create new skill", "create a SKILL.md for <X>", "scaffold a skill", "new skill entry", "register a skill", "신규 skill 추가", "스킬 만들어줘", "스킬 생성", "skill 만들어", or any request to add a new SKILL.md to this repo. ALWAYS invoke this skill BEFORE writing to skills/ or registry.yaml. DO NOT TRIGGER when: editing existing entry's content (just edit the file directly); migrating an existing entry (read AGENTS.md "Migrating from Existing Entries" first); only updating registry.yaml without creating a new SKILL.md.
This skill should be used after productive sessions to extract learnings and route them to appropriate Reusable Intelligence Infrastructure (RII) components. Use when corrections were made, format drift was fixed, new patterns emerged, or the user explicitly asks to "harvest learnings" or "capture session intelligence". Transforms one-time fixes into permanent organizational knowledge by implementing updates across multiple files.
Adopt, update, or verify a repository's Context-Driven Baseline through the public Roundfix Baseline Command. Use when preparing a repository for the write-prd/write-tasks/implement pipeline, changing its Baseline Profile, preserving existing instructions, restoring its Repository Skill Set, or confirming that Baseline-owned content is current
Creates production-grade, reusable skills that extend Claude's capabilities. This skill should be used when users want to create a new skill, improve an existing skill, or build domain-specific intelligence. Gathers context from codebase, conversation, and authentic sources before creating adaptable skills.
Install and update skills into the skill system from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, update existing skills, or install a skill from another repo (including private repos).
Converts Claude skills into ChatGPT Project format (prompt instructions + 1 knowledge file as .docx). Use when user mentions "convert to ChatGPT," "ChatGPT project," "export skill," "GPT instructions," "skill to prompt," or "skill to GPT."
Validates skills against production-level criteria with 9-category scoring. This skill should be used when reviewing, auditing, or improving skills to ensure quality standards. Evaluates structure, content, user interaction, documentation, domain standards, technical robustness, maintainability, zero-shot implementation, and reusability. Returns actionable validation report with scores and improvement recommendations.
Run the mandatory readiness gate for every ThreadWave Twitter/X workflow: require the update skill, confirm every release-index roster skill's individual version through GitHub, offer supported-version continuation, run approved skill and CLI updates as separate component-scoped flows through the canonical setup guide, verify the tw CLI and supported contracts, check the Chrome extension/setup relay, preserve the original request, and route sanitized failures to threadwave-error-support. Use before twitter-automation, twitter-agent, twitter-post, or twitter-reply, and for ThreadWave setup, update, readiness, dependency, or repair requests.
Check independently versioned ThreadWave Twitter/X skills against the authoritative GitHub release index without running the CLI or installing code. Use for latest-version checks, outdated or missing Twitter skills, version compatibility, flat skill installation diagnostics, or as the mandatory version authority called by threadwave-preflight. Approved installation belongs to threadwave-preflight.
Run five independent max-effort reviews (Fable 5.1, Opus 5.5, GPT-6 Astra, GPT-6.1 Sol, Muse Wonder), merge their findings, and fix only what the user approves. Manual only: run when a human types /total-review; agents must never invoke it on their own.
Run comprehensive triage-first verification pipeline with specialized skills. Detects scope, discovers toolchain, triages files to relevant skills, runs static analysis, invokes review skills in parallel, exercises the app, and produces a unified report. Supports interactive, report-only, and auto-fix modes.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.