The AI Skills Directory
We index, curate, and explain AI agent skills so you can find the right capability for your workflow β fast.
Browse SkillsWho we are
DiscoverAISkills is an independent directory built and maintained by a small team of AI practitioners and developers who use agent skills daily. We started this project because we found it genuinely hard to discover high-quality skills across the growing ecosystem of AI coding tools β so we built the resource we wished existed.
Our team has hands-on experience building and deploying AI agents with Claude Code, Cursor, Gemini CLI, and other tools. Every skill we index is reviewed for accuracy, and our descriptions are written to help you understand what a skill actually does β not just copy its README.
43,000+
Skills indexed
30+
Compatible AI tools
100%
Free to use
Open
Standard format
Our mission
AI agent skills are one of the most powerful β and most underused β features of modern AI tools. A single well-written skill can save hours of repetitive work, unlock new capabilities, and make your AI assistant genuinely useful for specialized tasks.
Our mission is to make the AI skills ecosystem discoverable, understandable, and accessible to everyone β from developers building production agents to individuals who just want their AI to do more. We believe the best skills should be easy to find, free to use, and open to all.
Curated Skills β Our Quality Standard
Not all skills are equal. We apply a rigorous evaluation framework called TRACEto identify the best skills in the ecosystem. Every skill that carries our "Curated" badge has been personally tested and verified by our team β we install it, run it in real workflows, and confirm it delivers on its promise.
Trust β the skill handles sensitive data responsibly and clearly declares its permissions and scope.
Reliability β it works consistently across environments with minimal failure, producing predictable and correct results.
Adaptability β it handles varied inputs gracefully and works across different platforms and contexts.
Convention β it follows the open Agent Skills standard, using proper frontmatter, documentation, and structure.
Effectiveness β it genuinely solves the problem it claims to solve and provides real value in day-to-day workflows.
When you see the Curated badge on a skill, it means our team has personally installed, tested, and validated it in real-world use. We stand behind every skill we mark.
What is an AI Skill?
A skill = a SKILL.md file that teaches an AI agent something new.
Without skills, an AI can only talk. With skills, it can act β following your workflows, using your tools, and completing real-world tasks on your behalf.
What a skill looks like
Every skill is a folder with a SKILL.md file at its root. The file has two parts: YAML frontmatter that tells the AI when to use the skill, and Markdown content with the actual instructions.
The name field becomes a /slash-command. The description helps the AI decide when to load the skill automatically β even without you explicitly invoking it.
How to install a skill
Installing a skill is as simple as placing the skill folder in the right directory. Skills can be personal (available across all your projects) or project-specific.
Personal skill (all projects)
Available in every session, across all your projects.
Project skill (this project only)
Scoped to the current project. Great for team-specific workflows.
After placing the skill, restart your AI client. You can then invoke it with /<skill-name> or let the AI discover it automatically based on context.
Core concepts
A skill is a folder of instructions
Each skill is a directory containing a SKILL.md file β a plain Markdown file with YAML frontmatter that tells the AI what the skill does, when to use it, and how to invoke it. No code required to create one.
Skills can include scripts and resources
Beyond SKILL.md, a skill folder can contain supporting scripts, templates, example outputs, and reference docs. The AI loads these on demand, giving it everything it needs to complete complex tasks reliably.
Skills work across AI platforms
The Agent Skills format is an open standard originally developed by Anthropic and now adopted by Claude Code, Cursor, Gemini CLI, OpenAI Codex, GitHub, VS Code, OpenClaw, and 30+ other tools. One skill, many agents.
Skills can be chained together
Agents become truly powerful when skills are combined. A research agent might chain a web-search skill β a summarization skill β a document-writing skill to produce a full report from a single prompt.
Example skills
Web Search
Let your agent search the internet in real time
Data Analysis
Analyze datasets and generate insights
Read and send emails on your behalf
Image Gen
Generate images from text prompts
File System
Read, write, and manage local files
API Connector
Call any REST or GraphQL API
Calendar
Schedule meetings and manage events
Code Runner
Execute code in a sandboxed environment
The Agent Skills open standard
The Agent Skills format was originally developed by Anthropic and released as an open standard at discoveraiskills.com. Because the format is platform-agnostic, a skill you write once works across every compatible tool β no modification needed.
This is why there is no real difference between a "Claude skill" and an "OpenClaw skill" β they are the same format. The platform just determines where you place the skill folder.
Ready to supercharge your AI?
Browse skills across every category β from productivity to coding to data analysis. All free, all open standard.