Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
OrchestKit doctor for health diagnostics across manifest integrity, hook configuration, skill validation, agent frontmatter, MCP server connectivity, CC version compatibility, and permission rules. Reports issues with severity levels and auto-remediation suggestions. Validates component counts, detects orphaned entries, and checks CC version matrix compliance. Use when diagnosing plugin health, troubleshooting configuration issues, or running pre-release checks.
Triggers when a user asks about, deploys, sets up, gets running, connects, checks, repairs, or upgrades GTM Agent (the Slack GTM agent that runs these skills on Vercel) for an organization, including questions before any setup about whether they could set it up, what it needs, or what it costs, with phrasings like “can I set up the Slack GTM agent”, “what would I need for the Slack agent? don't deploy anything”, “get gtm-agent running for Acme”, “deploy the GTM agent to our Slack”, “check the Acme deployment”, or “upgrade our GTM agent”. Load this skill for those questions instead of reading its files. Owns the agent's Vercel project, its Slack connector, and its access to the workspace repository. Not for the workspace, ICPs, personas, fit checks, or the workflows themselves, nor for taking the workflow project live or connecting it (gtm-workflow), which belong to gtm-workspace, gtm-icp, gtm-persona, gtm-qualify-prospects, and gtm-workflow
Find the map/territory gap BEFORE it gets expensive. Invoke when starting a non-trivial task in unfamiliar territory (new module, new technology, work the user has not done before), when acceptance criteria are "I'll know it when I see it", when the user struggles to describe what they want, or when a long-horizon task came back wrong and a blind retry is tempting. Not for typo-level fixes or tasks with fully objective acceptance criteria.
Use when building an agent system (orchestrator, pipeline) or metering LLM spend. Triggers: "agent system" / "агентная система", "LLM billing". Not for one LLM call or prompt wording (agent-harness), evals (agent-evals) or protocols (agent-interop). Opt out: "no agent layer" / «без агентного слоя».
Use when building or editing any AI feature in n8n: AI Agents, Text Classifier, Information Extractor, Sentiment Analysis, Summarization Chain, Basic LLM Chain, embeddings, vector stores, single one-shot LLM calls, or AI media generation (image / audio / video) via the native LangChain provider nodes. Triggers on any `@n8n/n8n-nodes-langchain.*` node, "agent", "chat assistant", "LLM with tools", "tool calling", "fromAi", "system prompt", "memory window", "structured output", "outputParser", "function calling", "RAG", "vector store", "embeddings", "classify with AI", "extract fields with LLM", "sentiment analysis", "summarize with LLM", "single LLM call", chat triggers with files, AI image / video / audio generation, or any multi-turn or one-shot LLM behavior.
Manage a stateful, run-directory-based proof project: continuation across runs, run-local source bookkeeping, manual GPT Pro handoff packages when a local attempt stalls, and an optional DeepSeek second opinion as additional evidence only. Use when the user asks for proof-run orchestration, a GPT Pro handoff, or cross-run proof continuation — use /proof-writer for ordinary proof drafting and /proof-checker for rigorous verification or submission acceptance.
Install Claude skills from GitHub repositories with automated security scanning. Triggers when users want to install skills from a GitHub URL, need to browse available skills in a repository, or want to safely add new skills to their Claude environment.
Strategic planning specialist for task breakdown, dependency management, timeline estimation, and resource allocation. Manual invocation only - expert at decomposing complex projects into manageable tasks, identifying dependencies, assessing risks, and creating actionable roadmaps. Use when starting complex projects, facing overwhelmed scope, needing structured approaches, or requiring systematic task management before implementation.
Use for Agently request-side setup and contracts: model settings, Prompt/input/output design, effect tuning, missing or redundant context, structured output, LongContent, auto_continue, response reuse, streaming, TTS/STT audio, session memory, embeddings, and retrieval within one request family. Use agently-design for cross-node data flow and model/Host ownership.
Entry point for MCP server development — interrogates the user about their use case, determines the right deployment model (remote HTTP, MCPB, local stdio), picks a tool-design pattern, and hands off to specialized skills. Triggers when the user asks to "build an MCP server", "create an MCP", "make an MCP integration", "wrap an API for Claude", "expose tools to Claude", "make an MCP app", or discusses building something with the Model Context Protocol.
PlantUML diagram quality on the agent-uml collaborative canvas — three tiers: rendering safety (syntax that prevents HTTP 400 blank canvas), conversation mechanics (when to push a version vs ask a question, what to write in the message parameter), and design effectiveness (decomposition thresholds, cross-diagram traceability, export readiness). Trigger whenever calling agent-uml MCP tools (design_create, diagram_upsert, design_feedback, design_export) — even when the task seems simple, since a missing `as alias` makes elements un-annotatable and a skinparam mismatch makes diagrams unreadable.
SUPERSEDED by /claudex-loop (formerly /crucible; adds a Phase 0 research/recon pass + rebuilt interview) — prefer that skill unless you explicitly want this one. Two-act plan hardening. ACT 1 (you ↔ Claude) — Claude interviews you relentlessly about a plan or design, one question at a time, recommending an answer for each and exploring the codebase when it can answer itself, until every branch of the decision tree is resolved. ACT 2 (Claude ↔ Codex) — Claude writes the locked plan to PLAN.md and OpenAI Codex adversarially reviews it in a read-only sandbox (VERDICT:APPROVED/REVISE); Claude revises and re-submits to the SAME Codex session until APPROVED or a MAX_ROUNDS cap, then you sign off before any code. Use when the user says "/grill-me-codex", "grill me then have codex review", "grill me and stress-test the plan", "interview me about this plan then get a second model on it", or is about to build something high-stakes (auth, schema, concurrency, migrations, payments) and wants both alignment AND a cross-model sanity check before implementation. Builds on Matt Pocock's grill-me (MIT). For the docs-aware variant use /grill-with-docs-codex; if you already have a plan and want only the Codex review use /codex-review. NOT for reviewing already-written code (use /codex:review) and NOT for trivial changes
Run a standardized benchmark to evaluate AI trading agent capabilities on the Manic Trade platform. Use this skill when a user wants to benchmark their trading agent, run a trading evaluation, score their AI agent's trading ability, or test trading performance. Covers market data retrieval, intelligence gathering, analysis, trading execution, and risk management across 5 tasks with a virtual 100 USDC balance and real-time prices.
Multi-angle codebase exploration spawning 3–5 parallel agents for code structure, data flow, architecture patterns, and health assessment. Generates ASCII visualizations, import graphs, and design pattern detection with cross-session memory storage. Use when exploring a repo, discovering architecture, onboarding to a new codebase, or analyzing design patterns.
Creates and orchestrates multi-agent pipelines on the iii engine. Use when building AI agent collaboration, agent orchestration, research/review/synthesis chains, or any system where specialized agents hand off work through queues and shared state.
Cổng sức khoẻ tổng / tuyến phòng thủ cuối của framework overstack — gõ /medic (hoặc `medic` ở CLI) để CHỨNG MINH hệ còn khoẻ trong MỘT lệnh: luật còn cắn (harness-doctor fire-drill), validator không lệch bản (drift), docs khớp đĩa, code compile, eval không regress, backstop git sống. Hub 1-tên — mô tả PHẠM VI thay vì nhớ subcommand (`medic` = all · `medic rules` · `medic docs eval`). Dùng trước commit/PR, sau khi pull/đổi máy, khi nghi một rule không chặn nữa, hoặc vừa thêm skill/rule/generator (medic tự báo nếu quên hàng rào). Trigger: /medic, "medic", "chạy medic", "hệ còn khoẻ không", "rà luật có cắn không", "cổng phòng thủ cuối", "health gate", "last line of defense".
Operate the Noodle CLI — explain commands, find flags, create/edit .noodle.toml config. Also covers writing skills for Noodle: the orders pipeline, task-type schedule fields, stage composition, and the orders-next.json schema. Use when running noodle commands, editing .noodle.toml, writing or updating a skill's schedule field, authoring new task-type skills, or asking "how to run noodle", "noodle help", "configure noodle".
Author a new Claude Code skill the house way, or refactor a fat skill into a lean SKILL.md + references/. Use when you want to "create a skill", "write a new skill", "turn a prompt or command into a skill", "split a skill into references", "trim a SKILL.md", or invoke /skills-create. The meta-skill — the skill that builds skills.
Entry point for 9Router—a local/remote AI gateway with OpenAI-compatible REST for chat, image, TTS, embeddings, web search, and web fetch. Use when the user mentions 9Router, NINEROUTER_URL, or wants AI without writing provider boilerplate. This skill covers setup and indexes capability skills; fetch the relevant capability SKILL.md from the URLs below when needed
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. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.
Post-merge workflow resolution. Verifies PR merge status, backfills synthesis metadata, force-resolves review statuses, transitions to completed, and cleans up worktrees/branches. Use when the user says 'cleanup', 'resolve workflow', 'mark as done', or runs /cleanup. Do NOT use before PRs are merged.
Expert guidance for creating Claude Code skills and slash commands. Use when working with SKILL.md files, authoring new skills, improving existing skills, creating slash commands, or understanding skill structure and best practices.
Error pattern analysis and troubleshooting for Claude Code sessions. Categorizes errors (network, auth, model, tool, memory, permission) with known resolution patterns, searches memory for prior occurrences, and suggests recovery steps. Delegates to debug-investigator agent for complex root cause analysis. Use when handling errors, fixing failures, or troubleshooting session issues.
Design a project-specific harness and create specialist agents and the skills they use. Use when the user asks to configure, build, or design a harness, engineer a harness, or create an agent team. Also use when building an automation system for a new domain or project, or restructuring or extending an existing harness. Use for operating or maintaining an existing harness, such as harness checks, audits, status reviews, or agent/skill synchronization. Use the harness:evolve skill to review execution results and incorporate feedback.
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).
Cross-session memory for AI. ALWAYS load this skill at session start to read SOUL.md and USER.md. This skill should be loaded for EVERY conversation to maintain continuity. Also triggers on "save memory", "end session", "update memory" for saving. Handles session history creation, memory consolidation, and USER.md/SOUL.md updates. Memory stored in ~/.learnwy/ai/memory/.
Writes or overwrites WORK-REPORT.md at the working tree root - session self-report (goal, problem/solution, follow-ups, suggested skills, abandoned) a reviewer or later session reads instead of re-deriving intent from the diff. Use for "work report", "checkpoint this", "hand off", "write up what we did". Offer proactively - asking first - when work is finished: "done", "ready to commit", "wrapping up", handoff to another agent.
Use for Agently runtime extensions: AgentExecution plugins and final review/artifact policies, Actions and MCP, Shell/BashExecutor and PowerShell permissions, ExecutionResource lifecycle, TaskWorkspace, RecordStore, auto_func and KeyWaiter, FastAPIHelper, or optional DevTools observation.
CodeStable documentation wrap-up. Triggers: end of a phase, organizing docs, syncing memory, /sync, /neat; performs a global knowledge base hygiene check and four-layer sync. Do not use for writing a single developer/user guide or API reference (cs-docs), fixing bugs (cs-issue), implementing features (cs-feat), or repository initialization (cs-onboard).
Autonomously achieve a goal. Use when the endpoint, acceptance criteria, or budget is clear; when iterating until completion; or when starting work after “grill me.” Don’t use for planning and implementing a single feature (cs-feat), end-to-end breakdown of a large request (cs-epic), bug diagnosis and fixes (cs-issue), behavior-preserving refactors (cs-refactor), external documentation (cs-docs), or pure design, roadmaps, brainstorming, or audits without an implementation endpoint.
Review failed Exarchos MCP tool calls from the current session, diagnose root causes, and categorize into code bug, documentation issue, or user error. Use when the user says 'dogfood', 'review failures', 'what went wrong', 'triage errors', or runs /dogfood. Scopes exclusively to Exarchos tools (exarchos_workflow, exarchos_event, exarchos_orchestrate, exarchos_view, exarchos_sync). Do NOT use for debugging application code or non-Exarchos tool failures.
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.