Awesome AI Agent Skills
103 ready-to-use AI agent skills for Claude Code, OpenAI Codex, Gemini CLI, Cursor, GitHub Copilot, Windsurf, and other tools that support the open Agent Skills standard.
Every skill is a complete, self-contained SKILL.md workflow for agent engineering, MCP, multi-agent systems, security, coding, research, data, DevOps, documents, marketing, sales, and more—not a link to another repository.
⭐ If this library saves you time, star the repository to bookmark it and help more agent builders discover it.
Quick start · Production demo · Starter packs · Browse 103 skills · Supported agents · Contribute
Quick Start: Install Agent Skills
Use the open-source Skills CLI to choose skills and install them for your detected agents:
npx skills add seb1n/awesome-ai-agent-skills
List all available skills without installing:
npx skills add seb1n/awesome-ai-agent-skills --list
Install one skill directly:
npx skills add seb1n/awesome-ai-agent-skills --skill agent-evaluation
Install several complementary skills by repeating --skill:
npx skills add seb1n/awesome-ai-agent-skills \
--skill agent-evaluation \
--skill agent-observability \
--skill human-in-the-loop
Starter Packs
| Pack |
Recommended skills |
| Production Agent Stack |
agent-evaluation, agent-observability, tool-schema-design, human-in-the-loop, prompt-injection-defense, agent-red-teaming |
| Multi-Agent Systems |
multi-agent-orchestration, mcp-server-building, agent-evaluation |
| Agent Security |
prompt-injection-defense, agent-red-teaming, skill-supply-chain-audit |
| Document Work |
pdf-processing, spreadsheet-analysis, presentation-creation |
Production Agent Stack Demo
Can a poisoned GitHub issue hijack a deployment agent? Run the offline demo with deterministic outcomes to compare a fragile agent boundary with a controlled stack built from six complementary skills:
python3 demos/production-agent-stack/run_demo.py --check
The committed fixtures move from 1/5 to 5/5 checks passed, with strict tool-schema, approval-policy, trust-boundary, trace-privacy, evaluation, and red-team evidence. The demo uses local mock effects only—no API keys, network requests, model calls, shell invocation, or real deployment.
Explore the Production Agent Stack demo →
What Are Agent Skills?
Agent Skills are reusable instruction packages that teach an AI agent when and how to complete a task. Each skill follows the open Agent Skills standard: a SKILL.md file with activation metadata and step-by-step guidance, plus optional scripts, references, and assets. Agents load a skill only when it is relevant, keeping context focused.
Why This Library?
Unlike collections that send you to external repositories, Awesome AI Agent Skills keeps every workflow in one inspectable, installable library.
- Ready to use: Install a single skill, a starter pack, or choose from the complete catalog.
- Cross-agent compatible: Use the same skills with Claude Code, OpenAI Codex, Gemini CLI, Cursor, GitHub Copilot, Windsurf, and other compatible tools.
- Validated structure: CI checks names, activation descriptions, links, duplicates, catalog synchronization, skill count, and bundled Python syntax.
- Progressive disclosure: Deeper skills bundle focused scripts, references, and assets without loading everything into context.
- Broad real-world coverage: Build agents for engineering, security, data, research, documents, compliance, finance, marketing, sales, and customer success.
- Open source: The library is MIT licensed and open to review, adaptation, and contribution.
Supported Agents
The Skills CLI detects supported agents automatically. For manual installation, copy a skill directory into the appropriate project or global path:
| Platform |
Project Path |
Global Path |
| Claude Code |
.claude/skills/ |
~/.claude/skills/ |
| OpenAI Codex |
.agents/skills/ |
~/.agents/skills/ |
| Gemini CLI |
.gemini/skills/ |
~/.gemini/skills/ |
| Cursor |
.cursor/skills/ |
~/.cursor/skills/ |
| GitHub Copilot |
.github/skills/ |
~/.copilot/skills/ |
| Windsurf |
.windsurf/skills/ |
~/.codeium/windsurf/skills/ |
Skills Index
Browse all 103 skills by category, or use the full searchable index below.
AI/ML Operations
Agent Engineering
| Skill |
Description |
| agent-evaluation |
Design reproducible evaluations, graders, baselines, and release gates for AI agents |
| agent-observability |
Instrument agent traces, metrics, cost attribution, dashboards, and investigations |
| human-in-the-loop |
Design auditable approval gates, escalation paths, and safe state transitions |
| mcp-server-building |
Build and verify secure, interoperable Model Context Protocol servers |
| multi-agent-orchestration |
Coordinate bounded multi-agent workflows, handoffs, shared state, and recovery |
| tool-schema-design |
Design precise model-facing tool schemas with explicit side effects and safe defaults |
Agent Security
| Skill |
Description |
| agent-red-teaming |
Plan authorized adversarial tests for agent behavior, tools, memory, and controls |
| prompt-injection-defense |
Defend agent workflows against direct and indirect prompt injection |
| skill-supply-chain-audit |
Audit skill bundles for malicious instructions, unsafe scripts, provenance, and excessive permissions |
API & Integration
Code & Development
| Skill |
Description |
| code-documentation |
Generate comprehensive code documentation |
| code-review |
Review code for bugs, style violations, and performance issues |
| debugging |
Identify and fix errors with root cause analysis |
| refactoring |
Improve code quality and maintainability |
| testing |
Write and run unit, integration, and end-to-end tests |
| version-control |
Manage Git workflows, branching, and merging |
Communication
Customer Success
Context Engineering
Data & Analytics
Database
Design & UI/UX
Documents & Files
| Skill |
Description |
| pdf-processing |
Extract, transform, create, and visually verify PDF documents safely |
| spreadsheet-analysis |
Analyze and modify spreadsheets while preserving formulas, structure, and auditability |
Finance & Accounting
DevOps & Infrastructure
Legal & Compliance
Marketing & SEO
Productivity & Workflow
Research & Knowledge
Sales
Security
Writing & Content
Each skill follows the Agent Skills open standard. A skill is a directory containing a SKILL.md file with YAML frontmatter and Markdown instructions:
skill-name/
├── SKILL.md # Required: Instructions and metadata
├── agents/ # Optional: Product-specific UI metadata
├── scripts/ # Optional: Executable scripts
├── references/ # Optional: Reference documentation
└── assets/ # Optional: Templates and resources
The SKILL.md file structure:
---
name: skill-name
description: What the skill does. Use when the user provides these inputs or requests this workflow.
---
# Skill Name
Instructions, workflows, and examples for the agent.
Quality and Validation
Run the zero-dependency catalog validator locally:
python3 scripts/validate_skills.py
The same validation runs in CI. New and materially revised skills must also test one normal prompt, one edge or failure prompt, and every executable script. See CONTRIBUTING.md for the complete quality and safety checklist.
Contributing
We welcome contributions from the community. Whether you want to add a new skill, improve an existing one, or fix a bug, please see our CONTRIBUTING.md for guidelines.
License
This project is licensed under the MIT License. See the LICENSE file for details.
评论