durable-agent-workflows

Reviews a pull request or local branch diff like a senior engineer, covering correctness, security, performance, naming, and missing tests, and produces prioritized, actionable comments (optionally posted to GitHub). Use when the user asks to review a PR, a diff, a branch, or "check my changes before I push".

Designs and writes new Agent Skills (SKILL.md) from a described workflow. It produces a precise trigger description, a step-by-step workflow, rules, an output format, optional helper scripts, and a README, then checks the new skill before handing it over. Use when the user wants to create, write, improve, or package an agent skill, SKILL.md, Claude skill, or reusable agent workflow.

Researches and writes SEO-optimized, genuinely helpful blog articles, covering keyword and search-intent analysis, competitor outline gaps, title and meta description, headings, FAQ, internal links, and schema, in the brand's voice and following Google's helpful-content and E-E-A-T principles. Use when the user asks to write a blog post or article, do SEO content, or rank for a keyword.

Converts a screenshot, mockup, or Figma export into a production-ready UI component that reuses the project's existing design system (React/Next, Vue, Svelte, React Native, Flutter, or SwiftUI with Tailwind, CSS modules, or a component library), so the result is responsive and accessible. Use when the user shares an image of a UI and asks to build, implement, or recreate it in code.

Roasts a GitHub profile with humor, then gives genuinely useful advice. It analyzes public repos, stars, languages, missing descriptions and licenses, abandoned projects, lazy commit messages, and activity with a data script, and produces a shareable scorecard and a glow-up plan. Use when the user asks to roast, rate, review, or improve a GitHub profile, or asks "what does my GitHub say about me".

Rewrites a repository's README into a polished, star-worthy page with badges, a hero tagline, demo section, quickstart, features, usage, and contributing info. Use when the user asks to improve, rewrite, polish, or "glow up" a README, or when a repo has a missing or weak README.

Scans an AI agent's instruction and tool-config files (CLAUDE.md, AGENTS.md, skills, slash commands, subagents, Cursor and Cline rules, hooks, MCP server configs) for prompt injection, hidden Unicode or HTML-comment instructions, exfiltration commands, and risky permissions or unpinned MCP packages, then recommends what to remove or quarantine. Use when the user installs third-party skills or plugins, clones an unfamiliar repo, adds an MCP server, or asks whether their agent setup is safe to trust.

Writes launch content for an open-source project or feature, including an X/Twitter thread, LinkedIn post, Show HN post, Reddit post, Product Hunt tagline, and dev.to article outline, based on the actual repo. Use when the user asks to announce, launch, promote, or write a post or thread about a repo, release, or feature.

Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.

Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.

Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes: Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/spec asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".

Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.

This skill should be used when the user wants to "write agent code", "build an agent with ADK", "add a tool", "create a callback", "define an agent", "use state management" — in a project that needs ADK (Agent Development Kit) API patterns and code examples. It provides a quick reference for agent types, tool definitions, orchestration patterns, callbacks, state management, the graph Workflow API, and reference recipes to study. Do NOT use for scaffolding (use google-agents-cli-scaffold) or deployment (use google-agents-cli-deploy).

Set up Claude Code hooks to block dangerous git commands (push, reset --hard, clean, branch -D, etc.) before they execute. Use when user wants to prevent destructive git operations, add git safety hooks, or block git push/reset in Claude Code.

Generate AI music on RunComfy via the `runcomfy` CLI — a smart router across the music-model catalog. Routes to ElevenLabs AI Music Generation (premium 44.1 kHz stereo vocal tracks, 5 s–5 min, $0.0083/s) and ACE Step / ACE Step 1.5 (StepFun-AI open-weights, tag-driven composition, multilingual lyrics, $0.0002–0.0003/s, ~27× cheaper), plus ACE Step audio-inpaint (regenerate a time range inside an existing track) and ACE Step audio-outpaint (extend a track before or after). Picks the right model for the user's actual intent — premium vocal hook, cheap background music library, multilingual pop song, repair a bad chorus, lengthen a 30 s draft into a 2 min cut — and ships each model's documented prompting patterns plus the minimal `runcomfy run` invoke. Triggers on "generate music", "make a song", "AI music", "background music", "instrumental track", "soundtrack", "jingle", "theme music", "royalty-free music", "compose", "music with lyrics", "extend music", "fix this song", "inpaint music", or any explicit ask to generate or edit music.

创建 lark-cli 的自定义 Skill。当用户需要把飞书 API 操作封装成可复用的 Skill（包装原子 API 或编排多步流程）时使用。