Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Coordinate multiple agents on complex game development features using battle-tested spawn recipes. Triggers on: "multi-agent coordination", "team orchestration", "feature handoff", "cross-domain collaboration", or when a task spans multiple specialist domains (combat system, narrative feature, UI overhaul, performance pass, launch prep). Do NOT invoke for single-domain tasks -- route directly to the appropriate specialist. Part of the AlterLab GameForge collection.
Set up Claude Code hooks to block dangerous git commands (push, reset --hard, clean, branch -D, etc.) before they execute. Use when user wants to prevent destructive git operations, add git safety hooks, or block git push/reset in Claude Code.
Use this skill when an approved task document already exists and the user wants code or content changes implemented. Reads docs/task/*.md, follows the implementation steps, updates TASKS.md, and can run the approved auto chain from implementation onward.
Educational guide on best practices for creating implementation plans that prevent drift. Covers style anchors, task sizing, TDD requirements, affirmative instructions, drift handling, and quality gates. Use when creating or improving implementation plans to ensure they follow proven patterns.
Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most one PR of proven corrections. Use for /maintain-verification-skill or "audit the verify skill".
Compatibility router for the shared optimization knowledge base and the language-specific optimization catalog skills. Use when: (1) selecting which optimization catalog skill to load, (2) the implementation language is not fixed yet, (3) a workflow still references the legacy optimization-catalog skill name, (4) deciding whether a finding is shared or language-specific, (5) updating the generalized knowledge-base structure.
AI-assisted pair programming with multiple modes (driver/navigator/switch), real-time verification, quality monitoring, and comprehensive testing. Supports TDD, debugging, refactoring, and learning sessions. Features automatic role switching, continuous code review, security scanning, and performance optimization with truth-score verification.
Pi terminal agent workspace — install, /commands, sessions, compaction, settings.json, AGENTS.md, skill discovery, precedence, session JSONL, /tree, providers/auth, models.json, credential resolution, platform setup, editor (@file, Tab, images, !bash), message queue (steering, follow-up), built-in tools (read/bash/edit/write/grep/find/ls, --tools, --no-tools), print mode (-p, piped stdin), SYSTEM.md/APPEND_SYSTEM.md, CLI flags (-c, -r, --session, --fork, --no-session), @files args, resource flags (--no-extensions, --no-skills), model shorthand (provider/id, name:thinking), all slash commands, env vars, philosophy (no MCP/sub-agents/plan-mode by design). Use for pi config, skills, sessions, compaction, providers, models, tmux, Termux, Windows, tools, modes, system prompt, message queue, or "does pi support X".
Use when starting any implementation task, feature request, bug fix, or refactoring work. Triggers on /plan command, before any code is written, when requirements need structured analysis, or when transitioning from brainstorming to implementation. Forces question-asking, approach comparison, and explicit approval before any code.
Three orthogonal analytical operations (WHERE/WHEN/WHY) + cross-operation synthesis. Each operation attacks the problem from a fundamentally different angle. The disagreements between the three ARE the valuable output. Works on any domain — code, business, strategy, design, text.
Analyzes conversation transcripts to extract user corrections, patterns, and preferences, then proposes skill improvements. Use this skill when users provide corrections, express preferences about code style, or when patterns emerge from successful approaches. Can be triggered manually with /reflect or automatically at session end when enabled.
Make a canister app discoverable to an AI agent from just its URL, including through ICP MCP. Covers publishing a /.well-known/ic-architecture manifest of your canisters — generated at deploy time from real canister IDs via the static-site recipe's presync hook and envsubst — plus exposing candid:service (interface), a getApiDoc query method (behavior), an optional OQL schema/execute data surface, and a /.well-known/ii-derivation-origin file (identity). Publishing that manifest opts the app into ICP MCP and is itself the operator's acceptance of the ICP MCP App Operator Terms, so confirm with the app's operator before deploying it. Use when making an app agent-ready or available through ICP MCP, for agent/service discovery, the well-known ic-architecture manifest, or generating that manifest at deploy time. Do NOT use for the sign-in flows that consume the derivation origin: use agent-web-identity for the agent/CLI web-identity link, or internet-identity for adding II login to a frontend.
Search Agent Skills in SkillsMP (an index of 1.6M+ public SKILL.md files covering Claude Code / Codex / ChatGPT), filtering by keyword, category, occupation, and language, and specifically dig up obscure Skills that are "well written but unknown". Use when the user says 找个 skill, 有没有现成的 skill, 搜一下 skill, skillsmp, skills 市场, agent skill 搜索, find a skill, search skills, discover skills. Also used to see which Skills already exist in a field and avoid reinventing the wheel. It also covers these colloquial phrasings: 这个功能有没有人做过, 别人写过没, 我想写个做 X 的 skill（动手前先搜）, 挖点冷门好用的, 有没有小众但写得好的, 这领域现在都有啥, 按星数排靠不靠谱, 只要中文的 skill.
Execute implementation plans by dispatching a fresh subagent per task with two-stage review (spec compliance then code quality). Use when you have an implementation plan with mostly independent tasks and want high-quality, fast iteration within a single session.
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.
Use when the user wants to build a CLI tool to automate browser interactions on a specific website using kimi-webbridge. Invoke when user says "create a CLI for X site", "build a tool to automate X", or wants to control a website programmatically via an AI agent.
Build agentic RAG systems with autonomous planning, reflection, and tool use. Use this skill when implementing agent-driven retrieval, query decomposition, iterative refinement, or multi-source RAG orchestration. Activate when: agentic RAG, A-RAG, autonomous retrieval, query planning, iterative RAG, multi-step retrieval, agent retrieval, RAG agent.
Audit the instructions an agent already carries — CLAUDE.md, AGENTS.md, skills, tool descriptions — for contradictions, over-constraint, and duplication, then propose a cut list. Use when an agent ignores its own instructions, when a CLAUDE.md has grown bloated, or when the user asks to audit or rightsize their agent context.
Get a rigorous read-only second opinion on finished work through the host's available reviewer capabilities, preferably from a different model provider. Falls back to a fresh same-provider agent only when necessary and labels the reduced independence. Every review includes a mandatory scope-fidelity check against the original requirements — unrequested additions, unrequested removals, and deleted or weakened tests. Use when asked to double-check, verify, cross-check, or get a second opinion, and before shipping high-stakes or complex work.
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.