Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Capture the current session's repeatable process into a reusable SKILL.md skill file. Use when the user wants to create a skill, save a workflow as a skill, turn a process into a reusable skill, or mentions "skillify", "create skill", "make a skill", "save as skill", "capture workflow", "turn this into a skill", "new skill", or wants to automate a repeatable process they just performed.
Customize Claude Code statusline. Use when: user says 'statusline', 'status line', 'customize statusline', 'modify statusline', 'statusline settings', 'statusline theme', 'change theme', 'color scheme', wants to add/remove/change segments (cost, git, model, context), switch color themes (catppuccin, dracula, nord), or asks what can be shown in the statusline.
Convert and browse session transcripts as HTML or Markdown. Supports Claude Code JSONL logs (auto-saved to ~/.claude/projects/) and GitHub Copilot CLI JSONL logs (auto-saved to ~/.copilot/session-state/*/events.jsonl). Auto-detects log source based on available directories and file format. Supports viewing the current session, a specific session by ID, agent background task output files, or all project sessions with optional date-range filtering.
Meta-skill for understanding and customizing Mindfold Trellis - the all-in-one AI workflow system for 9 AI coding platforms (Claude Code, Cursor, OpenCode, iFlow, Codex, Kilo, Kiro, Gemini CLI, Antigravity). This skill documents the ORIGINAL Trellis system design. When users customize their Trellis installation, modifications should be recorded in a project-local `trellis-local` skill, NOT in this meta-skill. Use this skill when: (1) understanding Trellis architecture, (2) customizing Trellis workflows, (3) adding commands/agents/hooks, (4) troubleshooting issues, or (5) adapting Trellis to specific projects.
Always-loaded discipline for every AppFollow toolkit skill. Read this before calling any AppFollow MCP tool. Enforces the preview→approval→confirm write ritual, the shared MCP error-code vocabulary, and the public-tools-only posture. Other skills reference this file rather than restating the rules.
Shared lifecycle contract for manual/auto control-flow plus optional ai-proxy overlay recognition in workflow skills: auto and overlay always revert to manual on completion or interruption; re-entry needs an explicit trigger; implicit continuation never re-activates them; batch orchestrators pass Stage-exit policy explicitly. Overlay maps to auto carrier + policy; charter stays in ai-proxy-discipline. Hosts that omit ai-proxy triggers ignore overlay. Referenced via frontmatter dependencies. Triggers — 「模式生命周期」「切换自动模式」「切换 ai-proxy」 / workflow mode lifecycle. Do NOT use as the proxy charter or as a third control-flow enum.
Add, list, and remove Bolna providers for telephony, LLM, speech-to-text, and text-to-speech services such as Twilio, Plivo, Exotel, OpenAI, Anthropic, Azure, ElevenLabs, Deepgram, Sarvam, Cartesia, Polly, and others. Use when the user wants to bring their own credentials.
Wire AG2's shipped tools into an `Agent` — both provider-native server-side tools (web search, web fetch, code execution, MCP, image generation, memory) and locally-executed common toolkits (filesystem, DuckDuckGo, Exa, Tavily, skills). Use when the user wants capabilities AG2 already ships rather than writing custom Python. For shell commands see `ag2-shell-tool`; for custom Python tools see `ag2-add-custom-tool`.
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to “open a website”, “fill out a form”, “click a button”, “take a screenshot”, “scrape data from a page”, “test this web app”, “login to a site”, “automate browser actions”, or any task requiring programmatic web interaction.
Run a structured multi-agent debate by spawning a panel of expert agents on any question, with convergence-aware iteration and typed synthesis output via the `agent-council` CLI. Use when a decision has genuine tradeoffs, high stakes, or hidden assumptions worth adversarial collaboration, or when confidence diagnostics matter more than a single recommendation. Compatible with any AI agent harness that supports agentskills.io skills (Claude Code, Cursor, Hermes Agent, OpenHands, etc.). Do not use for simple factual lookups, tasks with a clear correct answer, routine single-perspective work, or code execution and tool orchestration beyond debate.
Guide for coordinating multiple Claude Code instances as a team using Agent Teams. Covers TeammateTool operations, spawn backends, communication, task coordination, hooks, and orchestration patterns. Use when building multi-agent workflows requiring inter-agent communication.
Interactive tool that generates rules and instruction files for AI coding agents (Claude Code, Cursor, Windsurf or Devin Desktop, Copilot, Gemini or Antigravity, Codex, and others). Use when asked to set up agent rules, write CLAUDE.md or AGENTS.md style instructions, or standardize AI coding rules for a team.
Build LiveKit Agent backends in Python. Use this skill when creating voice AI agents, voice assistants, or any realtime AI application using LiveKit's Python Agents SDK (livekit-agents). Covers AgentSession, Agent class, function tools, STT/LLM/TTS models, turn detection, and multi-agent workflows.
Build backend AI with Vercel AI SDK v5/v6. Covers v6 beta (Agent abstraction, tool approval, reranking), v4→v5 migration (breaking changes), latest models (GPT-5/5.1, Claude 4.x, Gemini 2.5), Workers startup fix, and 12 error solutions (AI_APICallError, AI_NoObjectGeneratedError, streamText silent errors). Use when: implementing AI SDK v5/v6, migrating v4→v5, troubleshooting errors, fixing Workers startup issues, or updating to latest models.
Run the author's half of the document review loop without the user relaying each round — wait for the reviewer's <DOC>.revN.notes.md, answer it with <DOC>.rev(N+1).md, and repeat until the reviewer says ready for final, a round cap is hit, or only the user can move it forward. Use after rev1 exists, when the user invokes /answer-notes-loop <feature> [max-rounds] in the session that authored the plan, roadmap, technical or experience document. Its counterpart is review-doc-loop in a fresh session
Improve any existing Claude Code skill, CLAUDE.md, subagent definition, slash-command, or prompt markdown file by combining web-researched best-practice rewrites with autonomous Karpathy-style binary eval loops. Use when the user wants to "auto-research", "self-improve", "auto-improve", "optimize", "modernize", or "research-improve" any existing SKILL.md, CLAUDE.md, agent file under .claude/agents/, slash-command file under .claude/commands/, or any prompt markdown file. Triggers on phrases like "use auto-research to improve skill X", "auto-improve this CLAUDE.md", "research best practices and rewrite Y", or "self-research and optimize this agent". Researches authoritative sources (Anthropic docs, anthropic.com engineering blog, anthropics/skills GitHub) in parallel sub-agents, proposes a cited rewrite, runs an autonomous binary-assertion eval loop with git commit and reset bookkeeping for testable artifacts, and asks the user to confirm only at the end with a before-and-after score and full source list.
Provides AWS CloudFormation patterns for Amazon Bedrock resources including agents, knowledge bases, data sources, guardrails, prompts, flows, and inference profiles. Use when creating Bedrock agents with action groups, implementing RAG with knowledge bases, configuring vector stores, setting up content moderation guardrails, managing prompts, orchestrating workflows with flows, and configuring inference profiles for model optimization.
Automatically discover and recommend relevant Claude skills when users encounter tasks that could benefit from specialized capabilities. Use this skill proactively when detecting any of these patterns: (1) User mentions working with specific file formats (PDF, DOCX, Excel, images, etc.), (2) User describes repetitive or specialized tasks (data analysis, code review, deployment, testing, document processing), (3) User asks if there's a tool or capability for something, (4) User struggles with domain-specific work (React development, SQL queries, DevOps, content writing), (5) User mentions needing best practices or patterns for a technology, (6) Any situation where a specialized skill could save time or improve quality. Search using SkillsMP API (if configured), skills.sh leaderboard, or GitHub as fallback. Recommend 1-3 most relevant skills and offer to install via npx skills add.
Drive your own Claude Code session from Basecamp. Starts the local connector (basecamp connect), which listens to the account's event feed as a Basecamp agent and prints one line per trusted request: a mention, an assignment, a ping from its operator, a comment on a thread the agent follows, a boost on something it wrote. The connector boosts each one that asks for something the moment it arrives; this session puts that boost in its own words, picks the repo for the project, and hands the request to a background subagent that reads the context, does the work with your settings and tools, and replies as the agent. The session stays free to take the next request, and you can keep using it. Also connects an agent to this computer and manages its setup: the agent's credential (basecamp auth agent connect), who may give it work, and which projects it serves (basecamp connect setup), plus status and doctor. Use when asked to connect an agent, drive agents from Basecamp, watch Basecamp for an agent's mentions, start or stop the connector, or change who can give the agent work or which projects it serves.
BKN full-lifecycle orchestration entry point. Reads OpenBKN SDK / CLI operation references on demand. Whenever the task involves creating a BKN, extracting object relations from a PRD/document, generating or updating `.bkn`, data view binding, environment checks, test set generation, validation, or pushing, use bkn-creator first for workflow routing, stage gates, sub-skill orchestration, and result receipts. Applies to BKN create/read/update/delete/extract/copy/validate scenarios, skill package capability gap-filling / skill draft generation scenarios, and usage feedback inspection and improvement scenarios (scheduled task triggers, Agent conversation quality anomalies, feedback_brief input, continuous knowledge network optimization). Not for pure data semantic queries, which should be handled by data-semantic.
Use when the user has a fuzzy idea and wants to shape it into a concrete project spec before planning or building. Triggers on: \"brainstorm this\", \"I have an idea for...\", \"help me think through this project\", \"what should I build\", \"spec this out\". Also catches vague feature descriptions needing structured questioning to clarify scope. Does NOT write code, plan implementation, review strategy docs, or run strategy interviews \u2014 produces a WHAT/WHY spec through dialogue, not a HOW plan.
Analyze, reduce, and monitor Claude Code session context against a healthy "smart zone" (default 150k tokens, where quality stays high before context rot sets in). Use when the user wants to see how much of the smart zone is loaded at startup and the biggest consumers, decide whether to clean up (/clear vs /compact) or trim what loads every session (MCP servers, skills, memory, CLAUDE.md), pick/configure a status line, or set their own healthy window size. Triggers include "analyze startup context", "why is my context so big", "clean up context", "reduce context usage", "status line template", "configure statusline", "set healthy window", "smart zone size".
The keeper of the animal ecosystem. Not an animal themselves, but the one who understands the forest deeply enough to invite new creatures in. The druid communes with existing animals, walks through the grove to find the right name, and performs the ritual of creation. Use when creating new animal skills or gatherings for the Grove ecosystem.
Find reusable skills from the vercel-labs/skills registry by task keywords, evaluate fit, and suggest how to install or adapt them for MicroClaw. Use when users ask 'do we have a skill for X?', 'find skills for a given task', or want to discover existing skills instead of writing from scratch. Triggers on mentions of find skill, search skills, skill registry, or reuse skill.
Enable the GitHub CLI (`gh`) in Claude Code cloud sessions and GitHub Copilot coding agent environments. Use this skill when: (1) setting up a project so cloud AI agents can use `gh` for PRs, issues, and releases, (2) configuring setup scripts or SessionStart hooks for `gh` installation, (3) adding `copilot-setup-steps.yml` for GitHub Copilot agents, (4) troubleshooting `gh` auth failures in cloud sessions, or (5) configuring `GH_TOKEN` for headless environments. Triggers on: "enable gh", "github integration", "Claude Code cloud setup", "copilot setup steps", "gh auth in cloud", "gh not working in cloud", "setup script", or any request involving GitHub CLI access from cloud-based AI coding agents.
Recommend which GPT-5.6 model tier (Luna/Terra/Sol) and reasoning effort (Low..Ultra) fits a task, learn from the user's own successes and failures, diagnose quality problems, and report the live session's model and effort. Advisory — it proposes a grounded starting point and an escalation rule; the human decides and switches. Load when the user asks which model or effort to use, starts a substantial task without deliberately choosing a tier, reports a poor or unsatisfying result, wants to reduce cost without losing quality, or asks what model the session is running. Triggers — "which model", "which effort", "reasoning effort", "cut cost", "is a bigger model worth it", "do I need max/ultra", "result is bad", "quality dropped", "what model am I on", "record this combo".
Use when the user wants to add QQ platform support to the main branch of Hermes Agent stable, or explicitly mentions "add a QQ channel to hermes main", "install QQ support as a skill into Hermes", or "make stable Hermes support QQ and file sending". This skill patches the current repository into a version that supports QQ Bot, QQ file sending, QQ platform configuration, and toolset wiring.
Guide to creating and managing Claude Code hooks. Implementing PreToolUse, PostToolUse, Stop, and other hooks. When to use: (1) blocking dangerous operations, (2) suggesting actions after a tool runs, (3) cleaning up at session end, (4) customizing Claude Code. Example triggers: "create a hook", "PreToolUse hook", "block dangerous operations", "suggest after commit", "Claude Code hooks".
Configure IndyKite Agent Gateway (IAG) in front of agent-to-agent (A2A) workflows or MCP servers, optionally with the self-hosted IndyKite Token Service (ITS) that mints the delegation token carried in `X-IK-Token` so multi-hop chains grow hop by hop. Use when wiring up A2A or MCP policy enforcement, modeling workflows in the IKG, configuring `JARVIS_*` / `config.yaml` (including `token_service`), running the Token Service, reading IAG audit records, or debugging IAG 401/403 responses.
Build and maintain a Karpathy-style LLM knowledge base — a self-compiling Obsidian markdown wiki where an Agent ingests raw sources, compiles cross-linked concept/entity/summary pages, answers queries against the corpus, lints the graph for health, and audits in-context human feedback filed from Obsidian or the local web viewer. Use when (1) scaffolding a new knowledge base for any research topic, (2) ingesting articles/papers/PDFs/web pages into raw/, (3) compiling or restructuring wiki articles from existing raw material, (4) answering questions against the wiki and filing durable answers back, (5) running lint passes for dead links / orphan pages / coverage gaps / audit shape, (6) processing human feedback from the audit/ directory and applying corrections. Not for general note-taking, daily journals, or non-wiki Obsidian use.
Split a large task across parallel sub-agents with clean role boundaries and a merge step. Use when the user asks to "parallelize", "use multiple agents", "fan out", or when a task has independent workstreams that bottleneck on one agent.
(NS) Generate and update execution-handoff.md for planned SDD versions — operational prompt, per-task status table, token totals, and time tracking in seconds. Use at the end of planning after all task-NNN-*.md files exist, when tasks exist but handoff is missing, or when updating task status during version implementation. Do NOT use for ad-hoc coding without a version lifecycle, version partitioning (ns-sdd-version-partitioner), or GitLab-only issue execution (ns-execution-gitlab-issue).
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.