发现 Skills

>-

>-

>-

>-

Runs systematic tests on Claude Code customizations — skills, agents, hooks, and commands. Executes sample queries and validates responses against expected behavior. Use when testing whether a customization works correctly or when running functional and regression tests.

Generates a structured handoff document for the current session. Captures what was accomplished, what remains, open questions, and context needed by the next session or developer. Writes to docs/ai_handoffs/.

Shared product-marketing context for growth-pack skills. Read this before writing copy, planning launches, doing CRO, pricing, SEO, social content, or campaign strategy. Builds a minimal context bundle from the offer, audience, positioning, proof, competitors, and current experiments so downstream marketing skills stop improvising from thin air.

Run PinchBench benchmarks to evaluate OpenClaw agent performance across real-world tasks. Use when testing model capabilities, comparing models, submitting benchmark results to the leaderboard, or checking how well your OpenClaw setup handles calendar, email, research, coding, and multi-step workflows.

Curated documentation reference for developers building with Pinecone. Contains links to official docs organized by topic and data format references. Use when writing Pinecone code, looking up API parameters, or needing the correct format for vectors or records.

Review GitHub pull requests using GitHub MCP with clear severity levels and actionable feedback.

Build a competitive or casual trading card game deck. Covers archetype selection, mana/energy curve analysis, win condition identification, meta-game positioning, and sideboard construction for Pokemon TCG, Magic: The Gathering, Flesh and Blood, and other TCGs. Use when building a new deck for a tournament format or casual play, adapting an existing deck to a changed meta-game, evaluating whether a new set warrants a deck change, or converting a deck concept into a tournament-ready list.

Create interactive maps, elevation profiles, and spatial visualizations from GPX tracks, waypoints, or route data using R (sf, leaflet, tmap) or Observable (D3, deck.gl). Covers data import, coordinate system handling, map styling, and export to HTML or image formats. Use when visualizing a planned or completed tour route on an interactive map, creating elevation profiles for hiking or cycling routes, overlaying waypoints and POIs on a basemap, or building a web-based trip dashboard.

Meegle OpenAPI for work items: create, get, update, list, search, and related operations.

Analyze React/Next.js performance issues, N+1 queries, bundle size, and Core Web Vitals. Use when optimizing components, checking for slow renders, reviewing data fetching patterns, or when the app feels sluggish.

**ANALYSIS SKILL** — Review Azure Architecture Center multitenant documentation for accuracy, structure, and style. Produces a categorized review report. WHEN: "review multitenant doc", "review AAC doc", "ISV architecture review", "SaaS documentation review", "multitenant doc review", "review approaches doc", "review service doc", "check AAC structure". INVOKES: microsoft-learn MCP, search, read. FOR SINGLE OPERATIONS: Use Microsoft Learn MCP directly.

Master core development paths — Frontend, Backend, Full Stack, and DevOps. An atomic-skill approach for learning sequences and technology stack recommendations.

Complete Redis installation guide for all platforms - Linux, macOS, Docker, and cloud deployments with configuration templates

Overview

For repos with a backend/ (or similar) directory, treat that as the app root. Create all API routes and library code under backend/src/app/... and backend/src/lib/... Never create src/ at the repository root for backend routes. Use before creating any API or lib file to prevent 404s and wrong-path creation.

This skill should be used when the user asks to 'pick up an issue' or 'work on an issue'.

Implement designs as production-ready HTML/CSS with pixel-level fidelity. Garry Tan's design-html skill — translates mockups, Figma specs, or design descriptions into clean, semantic, accessible HTML output.

bkend.ai file storage guide. File upload, download, presigned URLs, bucket management, and CDN integration with bkend.ai BaaS. Triggers: bkend storage, file upload, file download, presigned URL, bucket, CDN. Do NOT use for: auth (use $bkend-auth), data operations (use $bkend-data).

Build and manage persistent multi-agent teams in OpenClaw using a three-tier architecture (Main → Manager → Workers). Use this when a user wants multiple AIs to divide and collaborate on work, form a dedicated team, or delegate different tasks to different assistants. Typical user requests: “I want to build a team,” “I need an assistant just for coding/reviewing/writing,” “Help me handle multiple things at once,” or “I need more help — let different AIs do different things.” The system automates persistent team construction and management. Technical triggers: multi-agent, sub-agent, SOUL.md, Manager Agent, Worker Agent, multi-agent team, team collaboration, orchestration. Main acts as a relay, Manager coordinates worker agents, and Workers execute specialized tasks. Communication uses explicit sessionKey patterns and prevents Main from contacting workers directly to maintain clear boundaries and security.

MUI Base UI style guidelines for building headless React component libraries (formerly headless-ui-style). This skill should be used when creating unstyled UI components, compound components with render props, accessibility-first patterns, or component libraries that separate logic from styling. Extracted from the MUI Base UI codebase (github.com/mui/base-ui).