English | 中文
AgentUse Skills provide a set of reusable and standardized procedural knowledge for AI Agents.
We have deeply organized existing AI Skills, provided the most convenient installation tutorials, further enhanced and verified compatibility for mainstream Agent frameworks (e.g., Claude Code, Open Hands), ensuring that every skill is "ready-to-use" out of the box.
- Deep Curation: More than just a collection; we've standardized the functions, prompts, and scripts for each skill.
- Convenient Installation: One-click scripts and detailed docs to significantly lower the barrier to entry.
- Framework Optimization: Specialized tuning for different Agent frameworks, verified by rigorous automated testing.
- Progressive Loading: Efficient metadata discovery ensures support for massive skill libraries without overwhelming the context window.
- Agent-Ready Certification: A specialized program for software vendors to bridge traditional applications with the AI Agent ecosystem through standardized APIs and verified integrations.
AgentUse Skills supports the following AI Agent frameworks:
- Claude Code
- Codex
- Cursor
- OpenCode
- OpenClaw
- Qoder
Each framework provides its own skill loading mechanism. Please refer to the installation guides in each skill directory for framework-specific instructions.
Agent-Ready is a certification program ensuring skills and software work seamlessly with AI Agent platforms.
| Skill Name | Description | Status |
|---|---|---|
| skill-market | Automated discovery and installation of specialized agent skills from the Zerone Skill Market. | 👑 Certified |
| Obsidian | Knowledge management app with an official CLI (v1.12+) supporting semantic search, graph queries, and vault management — up to 70,000x fewer tokens than traditional MCP approaches. | 👑 Certified |
The skill library is being rapidly organized, and more high-quality skills will be available soon!
| Skill Name | Description | Status |
|---|---|---|
| agent-browser | A fast Rust-based headless browser automation CLI with Node.js fallback. Enables Agent navigation, interaction, and snapshots. | 🔍 Pending Verification |
| github | Official GitHub CLI integration for managing Issues, PRs, and monitoring CI/CD Action runs. | 🔍 Pending Verification |
| content-research-writer | A collaborative writing partner that helps research, outline, draft, and refine content with proper citations while preserving your voice. | ✅ Verified |
| deep-research | Autonomous multi-step deep research skill powered by Gemini Deep Research Agent for market analysis, literature reviews, due diligence, and more — producing detailed, cited reports. | ✅ Verified |
| humanizer | Remove signs of AI-generated writing from text to make it sound more natural and human-written. | ✅ Verified |
| tavily-search | AI-optimized web search via Tavily API. Returns concise, relevant results for AI agents. | ✅ Verified |
| multi-search-engine | Multi search engine integration with 17 engines (8 CN + 9 Global). Supports advanced search operators, time filters, and WolframAlpha knowledge queries. | 🔍 Pending Verification |
| Skill Name | Description | Status |
|---|---|---|
| ui-ux-pro-max-skill | AI-powered design intelligence with 67 UI styles, 96 color palettes, and 100 industry-specific reasoning rules. | ✅ Verified |
| imagen | Generate images using Google Gemini's image generation capabilities. Pure Python implementation with zero dependencies. | 🔍 Pending Verification |
| canvas-design | Create museum-quality visual art through deep design philosophies, emphasizing master-level craftsmanship and minimal text. | 🔍 Pending Verification |
| frontend-slides | Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files, with 12 curated visual styles. | ✅ Verified |
| Skill Name | Description | Status |
|---|---|---|
| agentuse-share | A standardized workflow for contributing new skills to this repository, including research, documentation, and installation guide generation. | ✅ Verified |
| skill-market | Automated discovery and installation of specialized agent skills from the Zerone Skill Market. | ✅ Verified |
| superpowers | Battle-tested 20+ core skills including TDD, debugging, and collaboration patterns. | ✅ Verified |
| playwright-skill | Complete browser automation with Playwright. Auto-detects dev servers, performs testing, screenshots, and UX validation. | 🔍 Pending Verification |
| prompt-engineering | Advanced prompt engineering techniques including Few-Shot, CoT, and persuasion principles to maximize Agent performance. | 🔍 Pending Verification |
| self-improving-agent | Captures learnings, errors, and corrections to enable continuous improvement. Promotes learnings to project memory. | 🔍 Pending Verification |
| proactive-agent | Transform AI agents into proactive partners that anticipate needs and survive context loss using WAL and Working Buffer protocols. | 🔍 Pending Verification |
| Skill Name | Description | Status |
|---|---|---|
| openclaw-config-guard | A safety-first maintenance skill for auditing and repairing OpenClaw configuration with deterministic validation, backups, rollback, and change reporting. | ✅ Verified |
If you'd like to contribute new skills or improve existing ones:
- Structural Specification: Each skill must follow the
SKILL.mdstandard and include necessaryscripts/andresources/. - Verification Flow: Ensure the skill passed verification in at least one mainstream framework.
- Submit PR: We welcome your contributions via Pull Requests!
| Feature | Skills (This Repo) | Prompts | Project Context |
|---|---|---|---|
| Reusability | Extremely High; cross-session/project | Low; usually requires rewriting | Medium; workspace-specific |
| Efficiency | Progressive loading; saves tokens | Low; grows with conversation | Medium; depends on indexing |
| Extensibility | Supports complex scripts/resources | Pure text instructions only | Depends on file mounts |
| Maintainability | Centralized version management | Scattered distribution | Fragmented |
We advocate for an AI-native contribution workflow. You can use AI Agents along with our agentuse-share skill to automate the process of adding new skills.
- Fast-Track: Simply provide a GitHub URL to your AI Agent and ask it to use the
agentuse-shareskill. - Manual: Follow our contribution guidelines for structure and bilingual requirements.
For detailed instructions, please see our Contributing Guide.
- Only install skills from trusted sources (like this repo).
- Read the instructions in
SKILL.mdcarefully before production deployment. - Follow the principle of least privilege, granting only necessary permissions to the Agent.
A huge thanks to all developers contributing to the AI ecosystem! Special thanks to the following open-source communities for their inspiration and contributions to the field of AI skills:
© 2026 AgentUse Team. Released under the GPL-3.0 License.