Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
A thin orchestration layer for embeddedskills: discovers projects in the current workspace, selects a build/flash/debug/observe backend, chains .embeddedskills/state.json, and aggregates the results of the underlying skills. Triggers when the user explicitly enters one of the commands: "一键构建烧录", "自动诊断", "串起 build -> flash -> debug -> observe", or explicitly invokes /workflow.
Design production-grade multi-agent workflows with clear pattern choice (sequential, parallel, hierarchical), handoff contracts, failure handling, and cost/context controls. Use when architecting a multi-step agent pipeline, choosing between single-agent vs multi-agent approaches, or refactoring an LLM workflow that suffers from context bloat or unreliable handoffs.
Reference for the Claude Code CLI itself (code.claude.com). Covers CLAUDE.md memory, settings.json (All settings, settings precedence, managed-settings.json), env-vars, model config, keybindings, statusline, terminal config, cli-reference (-p / --continue / --resume), sessions, checkpointing, worktrees, remote control, channels, deep links, and troubleshooting.
Models problems and trade-offs without writing code or specifications. Use when a workflow needs silent reasoning, or the user asks for “/culture”, “no writes”, “rubber-duck this”, or “these bugs feel connected”. Not for specs (/mold), code (/cook), or bug diagnosis and fixes (/pasteurize).
The entry point to all Bright Data skills—start here to guide the agent to the right skill for the task: web data and scraping, fetching or unblocking a page, search, browser automation, or the CLI, SDK and MCP surfaces. Use when Bright Data is not set up or stops authenticating—the CLI is missing (“bdata is not recognized”), login is needed or fails (“No API key found”), there is no Bright Data account yet and one must be created without a browser, a call is refused with 401, 407 or another auth error, an expected zone like cli_unlocker is missing, a key must be set in a project’s .env, CI, or a container (BRIGHTDATA_API_KEY), or the user asks to install Bright Data, log in, get started, or which skill fits their task. Authentication for every other Bright Data skill is handled here.
Use the public `orca` CLI to operate Orca-managed worktrees, folder contexts, terminals, repos, automations, worktree comments, and the browser embedded inside the Orca app. Use when the user says "$orca-cli", "use orca cli", "Orca worktree", "child worktree", "cardStatus", "spawn codex/claude in a worktree", "read/wait/send Orca terminal", "terminal send", "full handoff", "handover", "give this to another agent", "another worktree", "Orca browser", or "control the browser inside Orca". Prefer this over raw `git worktree`, ad hoc PTYs, Playwright, or Computer Use when the task touches Orca-managed state. Use Computer Use for browser windows, webviews, or desktop UI outside Orca's embedded browser.
AWS Bedrock AgentCore comprehensive expert for deploying and managing AI agents at scale. Use when working with any AgentCore service including Gateway, Runtime, Memory, Identity, Code Interpreter, Browser, Observability, Agent Registry, or Evaluations. Covers agent deployment, MCP tool integration, credential management, agent discovery, governance workflows, and automated quality assessment. Essential when user mentions AgentCore, agent runtime, agent registry, agent evaluation, MCP gateway, deploy agent, register MCP server, discover agents, evaluate agent quality, agent credentials, or wants to build, deploy, catalog, or monitor AI agents on AWS.
Ann — Convener for MEL/SRHR work. Use when Ane brings any analytical, evaluation, SRHR, or structured-output task. Ann classifies task complexity, queries the MEL Wiki, retrieves knowledge, creates an implementation plan (verifies with user for complex tasks), coordinates with Vi for execution, runs a 5-point quality gate, and delivers. General-purpose — not tied to any specific project.
Vi — HR Specialist and Execution Orchestrator for MEL/SRHR work. Receives an approved plan from Ann (or directly from Ane), designs the specialist roster, spawns specialists as subagents, reviews their outputs, compiles the final product, and returns it. General-purpose — invoked by Ann via Agent tool, or directly by Ane when a plan is already approved.
CodeStable is an AI coding workflow for serious software engineering that keeps humans in the loop. Rather than centering on Agent orchestration, it organizes requirements, architecture, features, issues, and historical decisions to make development driven by Codex and Claude controllable, traceable, and continuously evolvable.
Create or update a shared Apple design context document that other HIG skills use to tailor guidance. Use when the user says "set up my project context," "what platforms am I targeting," "configure HIG settings," or when starting a new Apple platform project. Also activates when other HIG skills need project context but none exists yet. This skill creates .claude/apple-design-context.md so that hig-foundations, hig-platforms, hig-components-*, hig-inputs, and hig-technologies can provide targeted advice without repetitive questions.
Design and ship production-ready MCP (Model Context Protocol) servers from OpenAPI contracts instead of hand-written tool wrappers. Python and TypeScript support, schema validation, safe evolution. Use when exposing an existing API as an MCP server, building tool integrations for Claude or Codex or Cursor, or scaffolding an MCP project from scratch.
Guides the agent through connecting an MCP client to the hosted Capawesome MCP server, which exposes always-current Capawesome documentation search and the Capawesome Cloud management API. Covers setup for Claude Code, Claude Desktop, Cursor, and VS Code, API token creation, toolset selection, read-only mode, secret handling, verification, and troubleshooting. Do not use for installing Capacitor plugins, migrating apps or plugins to a newer version, running Capawesome CLI commands, or MCP servers other than Capawesome.
Adversarially stress-test a formed plan or design until every load-bearing assumption is defended or the plan is rejected. Use when the user asks to be grilled, pressure-tested, or invokes `grill-me-thermonuclear`. NOT for generating option spaces or collaborative ideation before a formed plan exists; use brainstorming-thermonuclear. Also not for affirmation, light polish, or executing a settled plan.
Comprehensive guide for maintaining, debugging, and upgrading the memory-lancedb-pro OpenClaw plugin — an enhanced LanceDB-backed long-term memory system with hybrid retrieval (Vector + BM25), cross-encoder reranking, multi-scope isolation, noise filtering, adaptive retrieval, and a management CLI. Use this skill when: (1) developing new features or fixing bugs in memory-lancedb-pro, (2) modifying the retrieval pipeline (vector search, BM25, RRF fusion, reranking, scoring stages), (3) adding or changing embedding providers, (4) updating scope/access control logic, (5) modifying agent tools or CLI commands, (6) troubleshooting memory quality issues (noise, duplicates, low recall), (7) working on the JSONL session distillation pipeline, (8) migrating data between memory backends, or (9) understanding the plugin's architecture to plan enhancements.
Guide a first-time user through Pipefy AI toolkit setup (Cursor Marketplace plugin, hosted MCP, local install.sh, or Claude Code plugin). Use when the user asks to install Pipefy MCP, connect Claude/Cursor to Pipefy, run /pipefy:install, or set up mcp.pipefy.com. Do not use for day-to-day Pipefy workflows after MCP is already working.
Session wrap-up. Use when Ane types /wrap-up to close a session. Gathers git, errors, and pending actions, delivers a concise status report, then commits and pushes pending work, and offers opt-in post-commit follow-ups (skill improvements, skill-fit signals, learning capture). Pass "report" for a read-only status with no commit. Writes a continuation handoff prompt for the next session whenever named work is still open, one file per workstream so parallel sessions cannot overwrite each other; pass "continue" to force one.
Testing and benchmarking LLM agents including behavioral testing, capability assessment, reliability metrics, and production monitoring—where even top agents achieve less than 50% on real-world benchmarks. Use when: agent testing, agent evaluation, benchmark agents, agent reliability, test agent.
Run a quality, sanity and efficiency audit of Ane's MEL/SRHR system (Ann/Vi/Li/Researcher team, wiki, harness, three-repo architecture): surfaces drift, inconsistencies, hygiene issues and architectural ceilings, categorised by severity, as a Tier 1 working brief with sequencing. Does NOT auto-execute fixes. Use when Ane types /system-audit, /audit, /system-check, /system-health, or asks whether anything is broken or drifting; run before any /li curate and after agent-improvement work. Different from /test (harness only), /li lint (wiki only), and /grade-system (overall quality posture rather than specific issues).
Retune a skill corpus for a new model, measurement-first: mine the run archive for a baseline, establish a noise floor, audit the corpus adversarially, then cut in measured passes until a pre-registered bar clears. Requires a benchmark harness that can A/B two builds of the corpus; refuses without one.
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.