skills.sh
skills.sh is an open directory of reusable agent skills built by Vercel Labs. Developers browse thousands of skills across topics like React, Next.js, testing, and agent workflows, then install them with a single command into Claude Code, Cursor, Codex, Copilot, and other agent runtimes. The platform is free, open source, and includes security audits, packs, and an API.
What is skills.sh?
skills.sh is the Agent Skills Directory, an open ecosystem where developers discover and install reusable skills for AI agents. A skill is a self-contained behavior module that encodes workflows, best practices, and domain expertise in the SKILL.md format, so instead of prompt-engineering the same behavior repeatedly, developers attach the right procedural knowledge to their agents and let them load it when needed. The directory hosts thousands of skills from organizations and individuals, including official skills from Anthropic, Vercel Labs, Microsoft Azure, and community contributors, with topics spanning React, Next.js, design and UI, mobile, agent workflows, databases, testing, and marketing. Skills are installed with a single command, and the platform supports popular agent runtimes including Claude Code, Cursor, Codex, GitHub Copilot, Windsurf, Gemini, Cline, AMP, Antigravity, OpenClaw, Droid, Goose, Kilo, Kiro CLI, and OpenCode. The ecosystem includes skills packs for curated collections, security audits that review skills for safety, and official badges for verified publishers. A leaderboard tracks installs and activity, with top skills like find-skills from Vercel Labs exceeding three million installs. The platform is built and maintained by Vercel Labs, is fully open source on GitHub, and provides documentation, a CLI, and an API for developers who want to build on top of the directory. Because skills follow the open SKILL.md standard, they are portable across compatible tools and agent runtimes, and teams can standardize how their agents behave across projects. The directory's audit system provides a review layer for community-contributed content, and official skills from trusted publishers carry badges that signal verified quality. Skills can be organized into packs for specific workflows, such as frontend development or testing, and teams can maintain private conventions by publishing their own skills and installing them consistently across machines. The platform's API enables programmatic search and installation, making it possible to integrate the directory into custom agent tooling, CI pipelines, and onboarding flows. Because the underlying SKILL.md format is open and lightweight, skills remain readable by humans and machines alike, which keeps the ecosystem portable and avoids vendor lock-in as new agent runtimes emerge.
skills.sh Core Features
Single-command installation
Install any skill with one command like npx skillsupdate
Thousands of skills
Browse skills from Anthropic, Vercel Labs, Microsoft Azure, and community publishers
Agent runtime support
Works with Claude Code, Cursor, Codex, GitHub Copilot, Windsurf, Gemini, Cline, and more
Skills packs
Curated collections of related skills for specific workflows and domains
Security audits
Audited skills give developers confidence in third-party content
Open SKILL.md format
Skills are portable across compatible tools and agent runtimes
Leaderboard and analytics
Track installs and activity for trending and popular skills
Open source with API
Fully open source on GitHub with documentation, CLI, and API for builders
Who is skills.sh for?
skills.sh is designed for developers and teams who work with AI coding agents and want to standardize how those agents behave. Software engineers using Claude Code, Cursor, Codex, or Copilot can install community skills for frontend design, test-driven development, codebase architecture, and debugging to improve agent output quality. Team leads and platform engineers use the directory to distribute consistent skills across their whole organization, ensuring every agent follows the same conventions. AI agent developers building custom agents or agent frameworks can publish their own skills and contribute to the open ecosystem. Technical founders building products on top of agent skills can use the API and CLI to integrate the directory into their tools. Open-source maintainers publish skills for their libraries so agents can use them correctly. Engineering managers standardizing development practices use packs and official skills to encode team processes. Even intermediate developers new to AI agents benefit from installing curated skills like find-skills or tdd instead of writing long prompts from scratch every time. Open-source maintainers publish skills that teach agents how to contribute to their projects correctly, and documentation writers use skills to encode style and structure rules for AI-assisted authoring. Security-conscious teams rely on the audit badges when adopting third-party content. Recruiters and coding bootcamps introducing AI pair programming can standardize the agent setup learners use. Platform engineers evaluating agent infrastructure compare skills across runtimes using the directory's compatibility pages, and independent developers exploring AI agents find curated starting points through the leaderboard and trending lists instead of building prompts from scratch.
skills.sh Use Cases
Install the tdd skill to make your coding agent follow test-driven development workflows
Add frontend-design skills so Cursor generates interfaces that match your design system
Standardize codebase architecture reviews across your team's agents with shared skills
Publish your library's usage guidelines as a skill for agents working with your code
Bundle related skills into packs for onboarding new developers to agent workflows
Use the API to integrate the skill directory into your own agent tooling
Audit third-party skills before adopting them in production environments
Browse trending skills to discover new agent capabilities used by the community
skills.sh Pros and Cons
Pros
- Free and open source, built and maintained by Vercel Labs
- Standardized SKILL.md format makes skills portable across agent runtimes
- Large ecosystem with official skills from major labs and thousands of community contributions
Cons
- Skills vary in quality since most are community-contributed, so audits matter
- The directory itself is a registry, not a hosted agent runtime, so users need their own agent setup
- Newcomers may need to learn the SKILL.md format and CLI before getting full value
FAQ About skills.sh
skills.sh Pricing
Free and open source. No paid plans; the directory, CLI, docs, and API are freely accessible.
Check official pricingFree
Browse, install, and publish agent skills; CLI, docs, and API included
skills.sh Alternatives
Polsia
Polsia is an AI platform that runs companies on autopilot. Built on agent SDKs and MCP integrations, Polsia plans roadmaps, ships code, runs ads, replies to customers, and posts content around the clock without human intervention. Founded by Ben Cera and operated by Polsia, Inc., it is free to start with no credit card required and subscription billing via Stripe.
Sollo
Sollo is an all-in-one AI platform that combines AI agents, automation workflows, and business tools in a single command center. Users deploy specialized AI employees for marketing, sales, customer support, and operations that work around the clock. Built for solo entrepreneurs, Sollo replaces an entire AI stack with one platform and includes access to GPT, Claude, Gemini, and Sollo's own models.
Tasklet
Tasklet is an AI agents platform that lets you delegate real work to agents connected to your tools, automating workflows around the clock with no developer required. Agents draft emails, summarize Slack, manage calendars, run automations, and learn over time, with thousands of built-in integrations, direct API access, and MCP support. Free to start with 300 bonus credits daily, SOC 2 compliant, with paid plans from $25 per month.