Git branch lifecycle management — create isolated worktrees for feature work, then finish with structured merge/PR/cleanup options. Use when starting feature work that needs isolation or when implementation is complete and ready to integrate.
Complete code review lifecycle — verify before claiming done, request structured reviews, and handle feedback with technical rigor. Use when completing tasks, before commits/PRs, or when receiving review feedback.
Shared vocabulary for designing deep modules — module, interface, depth, seam, adapter, leverage, locality. Use when deciding where a seam goes, designing or improving a module's interface, judging whether an abstraction earns its keep, or when another skill needs this vocabulary (notably sunnydata-design when…
Multi-source deep research using firecrawl and exa MCPs. Searches the web, synthesizes findings, and delivers cited reports with source attribution. Use when the user wants thorough research on any topic with evidence and citations.
Full design workflow — explore intent and requirements, write implementation plans, execute with review checkpoints. Use before any creative work or multi-step implementation.
Docker containerization and deployment patterns — Dockerfile best practices, Compose for local dev, CI/CD pipelines, deployment strategies (Rolling/Blue-Green/Canary), health checks, and production readiness. Use when containerizing apps, setting up local dev environments, or planning deployments.
Use when the user asks for a plain-language explanation ("白話", "講重點", "所以呢", "太長"), or when answering a decision question (does this affect me / should I act) whose honest answer would otherwise be buried in file paths, IDs and grep counts. Translates a verified finding to the reader's decision layer without dropping…
Pressure-test before spending effort. Mode A grades the ambiguity in a request and asks only what would change the deliverable, declaring the rest as assumptions. Mode B attacks a formed idea, plan, selection or business decision along fixed lenses and converges to the missing evidence, today's first step, and a…
Turn a decision you cannot answer alone into a questionnaire for someone else to fill in. Use when an answer lives in a person rather than a document — a domain expert, a client, a colleague on another team — and guessing would record an assumption as fact.
Comprehensive security review — OWASP Top 10 classification, implementation checklists, and language-specific best practices. Use when implementing auth, handling user input, creating API endpoints, working with secrets, or doing security assessments.
Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI. Provides project context, component docs, and usage examples. Applies when working with shadcn/ui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for…
Risk-based test development workflow with unit, integration, and E2E patterns. Use Red-Green-Refactor where it improves feedback, and apply project-defined coverage and evidence thresholds when writing features, fixing bugs, or setting up test infrastructure.
Chart a map for work too big and too foggy for one session — name the destination, ticket the decisions you can state now, and resolve them one per session until the way is clear enough to plan and build. Use when a greenfield project or large feature cannot be held in one session and the questions themselves are not…
A broad development guide written for Windows and PowerShell. It covers software development workflow, AI-assisted work, quick prototypes, dependency records, troubleshooting, and personal project notes.
Design and build AI agents for any domain. Use when users: (1) ask to "create an agent", "build an assistant", or "design an AI system" (2) want to understand agent architecture, agentic patterns, or autonomous AI (3) need help with capabilities, subagents, planning, or skill mechanisms (4) ask about Claude Code…
Perform thorough code reviews with security, performance, and maintainability analysis. Use when user asks to review code, check for bugs, or audit a codebase.
Build MCP (Model Context Protocol) servers that give Claude new capabilities. Use when user wants to create an MCP server, add tools to Claude, or integrate external services.
Claude Code instructions for Zenobia000/VibeCoding_Workflow_Templates, covering project initialization template, initialization notice, pre-initialization rules, 🚀 自動化專案初始化模板 and 🤖 claude code 初始化工作流程.