Awesome Agent Skills

English | 中文

Repository: https://github.com/YangsonHung/awesome-agent-skills

A collection of AI Agent Skills that provide professional domain capabilities for intelligent assistants like Claude Code.

Features

• Modular Design - Load skills on demand
• Bilingual Support - Every skill is maintained as an English and Chinese pair
• Language Guardrails - English SKILL.md files stay English-only, except native language names in multi-lang-readme
• Rich Resources - Complete references and templates for each skill

Available Skills

Trigger Examples

linkedin:

• "send a LinkedIn connection request to this profile"
• "message this person on LinkedIn"
• "search LinkedIn for marketing managers and extract their profiles"
• "post this update to my LinkedIn" novel-writer:
• "help me write a sci-fi novel"
• "write the next chapter"
• "design a villain character"
• "build a cyberpunk world"

multi-lang-readme:

• "create bilingual README"
• "translate README to Japanese"
• "add German README version"

conversation-json-to-md:

• "convert this chat-export json to markdown files"
• "one conversation per md file"
• "keep only user and assistant Q/A"
• "normalize output headings after export"

topic-bookmarks-reorganizer:

• "reorganize this bookmarks export and keep only the AI folder"
• "analyze links under one topic folder and regroup them"
• "dedupe repeated bookmark URLs and export importable html"

wechat-theme-extractor:

• "extract a theme from this WeChat article URL"
• "generate a markdown-wechat-converter theme from this article"
• "write the extracted style into markdown-to-wechat.html"

weixin-article-to-obsidian:

• "save this Weixin article into my Obsidian vault with local images"
• "download this mp.weixin.qq.com article as Markdown"
• "convert this WeChat article to an Obsidian note and localize images"

mac-software-storage-cleanup:

• "audit installed software sizes on my Mac"
• "show safe cache and simulator cleanup candidates"
• "give me a macOS storage cleanup plan"

ui-layout-analyzer:

• "analyze this UI screenshot"
• "describe this interface layout"
• "what does this UI do"
• "output the layout and function description of this UI"

yuque-lakebook-export:

• "export this Yuque .lakebook to markdown"
• "convert my lakebook into an Obsidian folder"
• "batch convert multiple .lakebook files"
• "fix missing images and broken internal links from Yuque export"

tweetclaw-twitter-automation:

• "scrape tweets with TweetClaw"
• "search tweet replies for this post"
• "post this approved tweet through OpenClaw"
• "export followers and summarize the result"

git-weekly-report:

• "generate a weekly report from my git commits"
• "summarize what I did this week across my projects"
• "create a weekly report from recent git activity"
• "review commits by project from last week"

git-push-secondary-merge-primary:

• "commit and push dev, then merge it into main and push"
• "push the work branch then merge to master"
• "merge develop into main, push both branches, and switch back to develop"
• "finish the work branch and sync it to the primary branch"

git-pushing-fast:

• "commit and push these changes"
• "push this work to the current branch"
• "save this work to remote"
• "commit with a segmented body and push"

git-commit-only:

• "commit these changes without pushing"
• "create a local commit"
• "commit current changes only"
• "save this work locally as a commit"

frontend-quality-guardrails:

• "review this React component for long text and layout overflow"
• "fix the mobile overflow and alignment issues in this dashboard"
• "polish this form and verify responsive UI states"
• "check this table for truncation, wrapping, and browser layout pitfalls"

react-component-hook-guidelines:

• "review this React component for split boundaries"
• "split this god component into Hooks and presentation components"
• "refactor this page to reduce props and state coupling"
• "check whether this Hook should move logic into utils"

logo-design-board:

• "design a logo for this AI note-taking app"
• "use a designer board to create brand mark directions"
• "critique these logo concepts and choose the strongest one"
• "turn this brand brief into logo prompts and vector notes"

feature-doc-splitter:

• "split this rough feature doc into overview, frontend, and backend docs"
• "turn this initial product note into implementation documents"
• "read the codebase and write frontend/backend build docs"
• "add Mermaid flows and ask whether to reserve design placeholders"

routerbase-api-integration:

• "migrate this OpenAI SDK call to RouterBase"
• "set up RouterBase as an OpenAI-compatible endpoint"
• "write a safe RouterBase API integration example"

routerbase-model-routing:

• "choose RouterBase models for this workload"
• "design a fallback chain for RouterBase"
• "write cost-aware model routing rules"

routerbase-media-generation:

• "build a RouterBase image generation workflow"
• "handle RouterBase video generation polling"
• "add safety checks for generated media assets"

x-twitter-scraper:

• "search X posts for this keyword"
• "export followers for this account"
• "set up an X monitor with a webhook"
• "choose the right Xquik endpoint for this workflow" ax-extract-workflow:
• "how did we ship this feature"
• "extract workflow from this commit"
• "what made this PR work"
• "reconstruct the recipe behind yesterday's demo" tree-ring-memory:
• "recall what we decided before changing this repo"
• "remember this lesson with evidence"
• "audit project memory before this release"
• "redact or forget stale memory about this task" xquik-data:
• "collect X posts with Xquik and summarize the results"
• "design an Xquik API workflow for this monitor"
• "use the Xquik MCP server for this social data task"
• "turn these Xquik webhook events into a handoff report"

Quick Start

Installation

Install with one of the following commands:

npx skills add YangsonHung/awesome-agent-skills

bunx skills add YangsonHung/awesome-agent-skills

pnpm dlx skills add YangsonHung/awesome-agent-skills

List available skills in this repository:

npx skills add YangsonHung/awesome-agent-skills --list

bunx skills add YangsonHung/awesome-agent-skills --list

pnpm dlx skills add YangsonHung/awesome-agent-skills --list

Install a specific skill:

npx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export

bunx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export

pnpm dlx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export

Install a specific skill globally:

npx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export -g

Install a specific skill to a specific agent:

npx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export -a claude-code

Install a specific skill without confirmation prompts:

npx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export -y

Install a specific skill globally for a specific agent without prompts:

npx skills add YangsonHung/awesome-agent-skills --skill yuque-lakebook-export -g -a claude-code -y

Or install manually by copying the English skills only:

cp -r skills/en/* ~/.claude/skills/

More command usage examples:

https://github.com/vercel-labs/skills/blob/main/README.md

Usage

Once installed, the skills will be automatically available in Claude Code. Simply ask Claude to perform a task related to a skill, and it will use the appropriate skill automatically.

Contributing

Contributions are welcome! To add a new skill:

1. Fork this repository
2. Create both paired directories under skills/en/ and skills/zh-cn/
3. Include the following files:
   • SKILL.md - Main skill definition with frontmatter
   • README.md - Optional skill-level notes
   • scripts/ - Optional helper scripts
   • references/ - Optional reference materials
   • assets/ - Optional templates and resources
4. Ensure both language variants are added together and kept in sync
5. Keep English SKILL.md files free of Chinese/Han characters; add Chinese examples and localized wording only to the paired skills/zh-cn/ skill. The exception is skills/en/multi-lang-readme/SKILL.md, where language-switch examples may show native language names.
6. Keep frontmatter description concise because it is used as routing metadata for skill selection:
   • English descriptions must start with Use when ...
   • Chinese descriptions must use natural Chinese trigger wording, not the English phrase Use when
   • Do not mention a specific agent in descriptions; describe the user need or task
   • Put longer purpose, capabilities, outputs, and boundaries in ## Overview
7. Submit a pull request

Test SKILL.md

Before opening a PR, validate skill files locally. The validators check skill structure, bilingual counterparts, and the English SKILL.md language rule:

python3 scripts/validate_skills.py
node scripts/validate-skills.js

For CI-level checks (fail on warnings), run strict mode:

python3 scripts/validate_skills.py --strict
node scripts/validate-skills.js --strict

Both commands must exit with code 0.

Git Hooks

Enable the repository pre-commit hook locally:

git config core.hooksPath .githooks

The hook runs strict skill validators, including the English skill language check, before each commit.

Skill Format

---
name: skill-name
description: Use when performing a concise task-specific trigger condition.
---

# Skill Name

## Overview

Explain the skill's purpose, capabilities, outputs, and boundaries.

## When to Use

Use this skill when the user asks for:
- Concrete trigger example
- Another supported task

License

This project is licensed under the MIT License - see the LICENSE file for details.