发现 Skills

Work with RabbitMQ — integrate, configure, and automate. Use when asked to set up RabbitMQ, use the RabbitMQ API, integrate RabbitMQ into a project, troubleshoot RabbitMQ errors, or build RabbitMQ automation.

Implements secure, scalable user sessions that survive server restarts and horizontal scaling. Use when building any authenticated web application that needs to track logged-in users across requests. Triggers on: session, login, authentication state, cookie, user state, stay logged in, remember me, session fixation, concurrent sessions.

Generate comprehensive Chinese technical tutorial documents for vLLM features and modules. Produces deep-dive code-walkthrough documents with Mermaid architecture/flow diagrams, comparison tables, code snippets, and performance analysis. Output is saved as Markdown to the skill's outputs/ directory. TRIGGER when: user asks to learn about a vLLM feature, module, or subsystem (e.g., "I want to learn about vLLM's xxx", "Help me generate a tutorial for vLLM xxx", "generate a tutorial for vLLM's xxx feature", "vLLM xxx feature analysis"). DO NOT TRIGGER when: user asks about non-vLLM projects, or asks for simple Q&A without requesting a tutorial document.

Fast headless browser for QA testing and site dogfooding. Navigate any URL, interact with elements, verify page state, diff before/after actions, take annotated screenshots, check responsive layouts, test forms and uploads, handle dialogs, and assert element states. ~100ms per command. Use when you need to test a feature, verify a deployment, dogfood a user flow, or file a bug with evidence.

Create a new Claude Code plugin. Triggered by phrases such as "create plugin", "new plugin", "make a plugin", "add plugin", "plugin create", "bulk create plugins", "full create", etc. Supports both creating standalone plugins and bulk creation that includes skills, sub-agents, and hooks.

Checks Elastic Beanstalk environment status, health, and deployment state. Lists environments and opens them in browser or AWS Console. Use when user says "eb status", "environment status", "what's running", "is it deployed", "check environments", "list environments", "eb health", "eb list", "eb open", "eb console", "is it healthy", or "how is my beanstalk". For troubleshooting use troubleshoot skill. For logs use logs skill.

Applies TDD process, test quality criteria, and mock guidelines. Use when: writing unit tests, using mocks, or reviewing test quality.

Requirements gathering through structured interview. Use when starting a new project or feature and need to build a comprehensive specification before implementation. Produces a specification document through guided discovery.

Migrate Python projects from setup.py/setup.cfg to pyproject.toml for use with uv. Use when upgrading legacy Python packaging, converting setup.py to a modern pyproject.toml layout, setting up dependency groups for development/testing, and ensuring `uv run pytest` works correctly.

Provides patterns for extending Shopper with custom sidebar items, component overrides, event listeners, and domain features like stock, pricing, and media. Use when customizing Shopper behavior.

Internal sub-skill of phpunit-unit-test-writing orchestrator. Not user-facing — invoked only via Skill(test-writing:phpunit-unit-test-generation) from the orchestrator.

Git-aware undo by logical work unit (track, phase, or task)

Configure secure, high-performance connectivity between on-premises infrastructure and cloud platforms using VPNs and dedicated connections. Use when building hybrid cloud architectures, connecting data centers or branch offices to public cloud providers, or requiring reliable, low-latency links.

Project scaffolding with deployment via GitHub Actions → Jenkins → Docker Compose. Use when setting up new projects, CI/CD pipelines, infrastructure scaffolding, or Docker deployments.

Convert documents and files to Markdown using markitdown. Use when converting PDF, Word (.docx), PowerPoint (.pptx), Excel (.xlsx, .xls), HTML, CSV, JSON, XML, images (with EXIF/OCR), audio (with transcription), ZIP archives, YouTube URLs, or EPubs to Markdown format for LLM processing or text analysis.

Systematic code analysis with evidence collection

Rooms as pipeline nodes, exits as edges, objects as messages

I mine pixels for atoms. Reality is just compressed resources.

Where you've been is where you can go back to.

Design technical architecture from an approved PRD. Produce PES (Product Engineering Spec), Mermaid diagrams, and API contracts. Use when a PRD or structured requirement has been approved and needs to be translated into system design. Do NOT use for coding, testing, or when requirements are still vague — redirect to requirements-analyst first.

ビジネス・開発・デザイン3チーム横断の統合仕様パッケージを作成。段階的詳細化テンプレート（L0ビジョン→L1要件→L2チーム別詳細→L3受入基準）で共通認識を形成。コードは書かない。

Build terminal UIs, support CLI development, and integrate development tools (linters, test runners, build tools). Use when you need to design and implement command-line experiences. Language-agnostic with support for Node.js, Python, Go, and Rust.

AI 3D model generation agent. Generates text-to-3D and image-to-3D code (Python / JavaScript / OpenSCAD) using Meshy, Tripo, Hunyuan3D, Rodin, Sloyd, and Stability APIs. Handles game-pipeline integration, LOD generation, retopology, UV packing, and QC validation.

A multi-session parallel orchestrator using the Claude Code Agent Teams API. It launches and manages multiple Claude instances to execute tasks concurrently. Use when parallel work is required.