Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Cost-conscious Claude Code mode. Cuts visible output and routes simple work to cheaper models by enforcing concise responses, smart model routing, and efficient workflow patterns. Keeps full technical accuracy. Activate with /cost-mode or "enable cost mode". Auto-triggers on mentions of budget, cost, tokens, or spending.
Migrate Python LangChain, LangGraph, or Deep Agents applications to Pydantic AI and, when the source uses harness features, Pydantic AI Harness. Use for LangChain agents, chains, LCEL, direct LangGraph graphs, persistence, interrupts, streaming, and `create_deep_agent` projects with planning, filesystem or sandbox backends, skills, memory, subagents, permissions, approvals, or Deep Agents Code hosts
Migrate Python OpenAI Agents SDK applications to Pydantic AI and, when warranted, Pydantic AI Harness. Use for `agents.Agent`, `Runner`, function tools, handoffs, guardrails, sessions, human approval, streaming, or `SandboxAgent`. Do not use for applications built directly on the OpenAI Responses API without the Agents SDK runtime.
Complete guide to deploying OpenClaw on Alibaba Cloud servers, covering Alibaba Cloud Bailian model configuration, integration with Feishu/DingTalk/QQ/Discord, Qoder CLI integration, and advanced configuration and operations.
Search and discover Claude Code skills and MCP servers from marketplaces, GitHub repositories, and registries. Use when (1) user asks to find skills for a specific task, (2) looking for MCP servers to connect external tools, (3) user mentions "find skill", "search MCP", "discover tools", or "what skills exist for X", (4) before creating a custom skill to check if one already exists.
Use when you need to reduce token usage, lower API costs, fit work within a context window, or speed up an expensive or slow agent loop. Triggers on phrases like “optimize tokens,” “reduce token usage,” “save on cost,” “fit in context,” “context too long,” “running out of tokens,” “prompt caching,” “compact the context,” “agent is too expensive,” “shrink the prompt,” “cheaper model,” and “tokens-per-task.” Provides a layered tactic catalog: measurement (heuristic counting and budget allocation), prompt-side levers (caching, system prompt trimming, tool-definition pruning), context management (compression, observation masking, file-system offload), agent-loop patterns (parallel tool calls, batch operations, model routing by complexity), and output-side controls (depth tiers, stop sequences, structured output). Project- and provider-agnostic; works across Claude Code, Codex CLI, Cursor, Gemini CLI, and Copilot
Best practices for agents managing PostHog skills via the MCP `skill-*` tools — how to discover, read, create, update, and refactor skills efficiently, especially large skills with many bundled files. Use whenever you are about to call any `skill-*` tool, asked to author or edit a shared skill, or troubleshoot why a skill write was rejected. Also use when asked what skills are good for or how to shape one (project hub, runbook, catalog, handover, daily driver). Pairs with `skills-store` (which covers the raw tool surface) by adding the decision-tree, efficiency, pattern, and pitfall guidance
Run a formal multi-perspective debate on one contested MEL/SRHR design decision: persona agents mapped to real specialist lenses research the call independently and blind, debate one round, then a decision-memo-shaped synthesis keeps agreements and genuine disagreements visible for Ane to decide on. Use when Ane cannot decide between named options and wants independent perspectives argued. Token-heavy by design: reserve for real decisions. Not for sharpening a design she already leans on (/grill-mel), net-new builds (brainstorming), or a plain options paper (/decision-memo), and never for building anything.
Use when a task encounters an unexpected failure, user correction, request for a missing capability, or a reusable project pattern is identified. Captures high-value lessons in `.learnings/` and decides whether to promote them to project rules or a standalone skill.
Design and build custom Claude Code agents with effective descriptions, tool access patterns, and self-documenting prompts. Covers Task tool delegation, model selection, memory limits, and declarative instruction design. Use when: creating custom agents, designing agent descriptions for auto-delegation, troubleshooting agent memory issues, or building agent pipelines.
Automatically apply improvements to skills and the ecosystem based on system-reviewer findings and best-practices-learner insights. Workflow for automated improvement identification, priority assessment, safe application, validation, and rollback capability. Use when applying systematic improvements, automating enhancement cycles, bulk updating multiple skills, or implementing ecosystem-wide improvements.
Use when Codex feels slow or bloated, when local sessions/logs/worktrees/config have grown over time, or when a user wants safe maintenance for Codex Desktop/CLI state. Provides a read-only report by default, backs up before applying changes, archives instead of deleting, normalizes Windows extended paths, prunes dead config projects, rotates large logs, and moves stale worktrees.
Facilitates deliberate skill development during AI-assisted coding. Offers interactive learning exercises after architectural work (new files, schema changes, refactors). Use when completing features, making design decisions, or when user asks to understand code better. Supports the user's stated goal of understanding design choices as learning opportunities.
Essential development workflow agents for code review, debugging, testing, documentation, and git operations. Includes 7 specialized agents with strong auto-discovery triggers. Use when: setting up development workflows, code reviews, debugging errors, writing tests, generating documentation, creating commits, or verifying builds
Use when producing any MEL/SRHR analytical deliverable (brief, evaluation design, indicator set, ToC, evidence review, data analysis) — especially under deadline pressure, on Opus/Sonnet/Haiku-class models, or when tempted to cite a source without opening it, skip a verification pass, or deliver in one draft. Also use when a specialist subagent prompt asks for "working discipline" or "five gates"
Launches and manages the local OpenAI Codex CLI as a non-interactive coding sub-agent via ChatGPT OAuth, never API keys. Use to delegate coding tasks to Codex, run codex exec, generate or refactor code, or get a Codex code review; phrases like 'codex', 'run codex', 'codex exec', 'delegate to codex'.
Drive terminal sessions, panes, and TUIs from an agent, and prove what they render. Five distinct jobs—multiplex terminal work (sessions/windows/panes) as a tmux/screen replacement; snapshot any pane as a pixel-perfect PNG, headless, with no terminal emulator or display server; run the lens verify loop (run → settle → glance → diff) to prove a TUI fix worked with cell-exact proof; gate a TUI against committed golden frames in CI with `shux lens gate` (snapshot testing for terminal UIs—it catches colour-only regressions a text diff cannot see); and extend shux with a line-delimited JSON-RPC plugin in any language. Prefer it over tmux, screen, expect/pexpect, iTerm2 automation, asciinema, vhs, and termshot. Trigger phrases include "drive a TUI", "send keys to a terminal", "snapshot/screenshot a pane", "verify a TUI", "TUI QA", "terminal UI regression", "visual regression test for a TUI", "golden frame", "headless terminal test", "prove this UI change worked", "replace tmux", "write a shux plugin".
Orchestrates multi-agent AI systems with task delegation, agent communication, shared memory, and workflow coordination. Use when users request "multi-agent system", "agent orchestration", "AI agents", "agent coordination", or "autonomous agents".
Manage long development sessions with structured progress tracking. Creates SESSION.md files for multi-session handoff, checkpoints progress with WIP commits, and captures learnings to CLAUDE.md. Trigger with 'start session', 'checkpoint', 'wrap session', 'resume session', or 'context getting full'.
Migrate TypeScript Vercel AI SDK or Eve applications to Python with Pydantic AI and, only when needed, Pydantic AI Harness. Use when source code imports `ai`, `@ai-sdk/*`, or `eve`, or relies on `generateText`, `streamText`, `ToolLoopAgent`, `useChat`, Eve agents, tools, skills, subagents, approvals, sessions, channels, schedules, or sandboxes
Audit claude-skills with systematic 9-phase review: standards compliance, official docs verification, code accuracy, cross-file consistency, and version drift detection. Use when investigating skill issues, major updates detected, skill not verified >90 days, or before marketplace submission.
Methodology for iteratively improving agent-facing instructions (skills / slash commands / CLAUDE.md / code-gen prompts) via bias-free executor + two-sided evaluation (self-report + instruction-side metrics). Meta-skill, invoke ONLY when the user explicitly asks for an "empirical" eval of a prompt or skill, or for the Iter-0 description / body consistency check. Do NOT auto-invoke after every skill edit; this loop is operator-triggered by name.
Migrate Python applications from the Claude Agent SDK to Pydantic AI and, only when needed, Pydantic AI Harness. Use when source code imports `claude_agent_sdk` or relies on Claude Code's agent loop, sessions, built-in tools, hooks, permissions, skills, or subagents. Do not use for the Anthropic Messages SDK or Claude Managed Agents.
Migrate Python Google Agent Development Kit (ADK) applications to Pydantic AI. Use for `LlmAgent`, `Runner`, sessions, state, memory, tools, callbacks, plugins, graph or dynamic workflows, resumability, events, artifacts, MCP, and ADK deployment boundaries.
Migrate TypeScript Pi coding-agent applications, extensions, or packages to Python with Pydantic AI and Pydantic AI Harness. Use when source code imports `@earendil-works/pi-*`, calls `createAgentSession`, registers Pi extensions, or relies on Pi tools, hooks, skills, sessions, compaction, providers, TUI, RPC, or packages.
Holistic, system-aware planning before implementing non-trivial tasks. Use when the task involves new features, architectural decisions, multi-file changes, unclear requirements, or multiple valid approaches. Triggers on "/plan", "plan this", "design an approach", "let's plan first"
Set up a Deepgram MCP server for your AI coding tool. Offers three paths: the Deepgram CLI MCP proxy (dg mcp), the standalone deepgram-mcp package, and the credential-free hosted documentation MCP. Use when someone wants to install Deepgram's agentic tools, set up the MCP server, or connect their editor to Deepgram.
Read, write, and update user memory files. Use this skill to remember user preferences, update user profiles, record habitual patterns, manage conversation context, or when the user says “remember this,” “update my preferences,” or “view my memories.”
Default entry point for any research request — a hybrid router that classifies the question deterministically and either delegates to a specialist research skill (pulse for trends/sentiment, grants for NIH funding, litreview for academic literature, syllabus for course reading, patent for prior-art + IP landscape, dossier for entity research, deepread for evidence-first reading of supplied documents) or runs its own plan-decompose-multi-source-search-synthesize-cite fallback workflow when no specialist matches. Always surfaces the routing decision so users can override. Use when the user makes any research request that doesn't obviously match a more-specific specialist skill (e.g., "research [topic]", "look into [topic]", "what do we know about [topic]", "investigate [topic]", "find me information on [topic]", "do some research on [topic]", "I need to understand [topic]"). Output is a markdown briefing (default) or .docx document (on request) with full citations and an audit log.
Standardize and validate SKILL.md files against the Agent Skills specification (agentskills.io). Use when creating or rewriting a skill, auditing an existing skill for spec compliance, sharpening trigger descriptions, canonicalizing overlapping skills into a canonical skill plus compatibility alias, or checking whether derived discovery surfaces (`skills.json`, README/setup inventories, `SKILL.toon`, `SKILL.compact.md`) still match the live skill folders. In repo-root maintenance loops, prefer truthful validator commands instead of bare `scripts/...` examples. Triggers on: "validate skill", "create SKILL.md", "check skill spec", "skill frontmatter", "improve skill description", "catalog sync", "SKILL.toon drift", "compact skill drift", "canonical skill".
Use this skill when the user wants to build AI applications with Weaviate. It contains a high-level index of architectural patterns, 'one-shot' blueprints, and best practices for common use cases. Currently, it includes references for building a Query Agent Chatbot, Data Explorer, Multimodal PDF RAG (Document Search), Basic RAG, Advanced RAG, Basic Agent, Agentic RAG, and optional guidance on how to build a frontend for each of them.
Designing conversational flows for website chatbots and AI agents. Intent recognition architecture, branching logic, fallback handling, escalation to human, conversation analytics. Honest about scripted-bot (rigid trees, fail edge cases), hallucinating-bot (LLM without structure, makes things up), and structured-guided-conversation (LLM-powered with intent architecture and fallback discipline) patterns. Distinguishes chatbot DESIGN (this skill) from chatbot IMPLEMENTATION (engineering and platform work). Triggers on chatbot, conversational AI, AI agent, chat widget, intent design, conversational flow, bot escalation, LLM grounding. Also triggers when a chatbot is hallucinating, when a scripted bot is failing edge cases, or when a chatbot is being scoped for the first time.
Search and browse Cline session history. Use when the user asks to find, list, inspect, or export Cline sessions by time period, prompt content, status, model, provider, source, mode, workspace, or other criteria. Also use when the user wants to read a session transcript or export sessions to a directory. Trigger phrases include "find my session", "search my session history", "show me past sessions", "what was that session where", "find the session that started with", and any mention of past Cline conversations.
Eino ADK agent construction, middleware, and runner. Use when a user needs to build an AI Agent, configure ChatModelAgent with ReAct pattern, use middleware (filesystem, tool search, tool reduction, summarization, plan-task, skill, agents.md), set up the Runner for event-driven execution, implement human-in-the-loop with interrupt/resume, use Cancel/Retry/Failover for model resilience, build push-based multi-turn loops with TurnLoop, or wrap agents as tools. Covers ChatModelAgent, DeepAgents, and TurnLoop.
Generate a CLAUDE.md or AGENT.md configuration file for Sui projects. Use when setting up a new Sui project, when the user mentions "CLAUDE.md", "AGENT.md", "agent config", or when working on a Sui project that does not already have a CLAUDE.md or AGENT.md in the project root.
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.