agentskill.sh

What is agentskill.sh?

agentskill.sh is an AI agent skills directory and marketplace for finding and installing pre-built “skills” into multiple AI tools. It’s positioned as a single place to browse a large library of skills and install them with a command.

The directory covers skills for platforms including Claude Code, Cursor, Copilot, Windsurf, Zed, and other AI tools. Skills are organized by role (e.g., Product Management, Data & Analytics) and by platform, with individual skill pages showing quality and security audit information.

Key Features

• Large skills directory (e.g., “106,000+ skills”) across multiple AI tools, so you can search for functionality instead of building from scratch.
• One-command installation using a /learn/learn @owner/skill-name syntax shown on the site, intended to streamline adding a skill.
• Browsing and discovery by role and by platform (e.g., “Skills by Role” and “Browse all Skills by Platform”).
• Curated bundles of skills (example shown: “Claude Skills”) that group multiple installable skills under a single set.
• Skill page metadata including “Quality Score” and “Security Score” with audit notes (when available), plus a “tool wrapper pattern” note on many entries.

How to Use agentskill.sh

1. Browse the directory by platform (e.g., Claude Code, Cursor, Copilot, Windsurf, Zed) or by role categories such as Engineering or Marketing.
2. Open a specific skill page to review what it does and any listed “Quality Score”/“Security Score” and audit information.
3. Install the skill using the site’s documented command format: /learn/learn @owner/skill-name.
4. If you want multiple related capabilities at once, start from a curated bundle (for example, “Claude Skills”) and install skills from that set.

Use Cases

• Install a tool wrapper for offline text-to-speech: find the sherpa-onnx-tts4.0 skill entry and use it to enable local audio generation from text via the referenced sherpa-onnx CLI.
• Manage Feishu knowledge bases: use the feishu-wiki skill to create, move, and edit wiki pages, then pair it with other Feishu-related skills (such as permissions or drive operations) from the directory.
• Automate GitHub issue triage: install the gh-issues skill to fetch, fix, and monitor pull requests using sub-agents as described on the skill listing.
• Route plain-language coding requests to an agent runtime: use the acp-router skill to direct requests into OpenClaw ACP runtime sessions or acpx-driven sessions.
• Publish or moderate content via the X API: install the xurl skill to perform authenticated interactions for posting, replying, and managing content.

FAQ

• How do I install a skill from agentskill.sh? The site shows installing a skill with /learn/learn @owner/skill-name.

• What does “Quality Score” and “Security Score” mean on skill pages? Individual skill entries display these scores and, in some cases, include the date of audit and whether issues were detected (as shown on example listings). The exact scoring methodology is not described on the provided page content.

• Can I browse skills by role instead of by platform? Yes. The directory includes categories grouped by roles such as Product Management, Data & Analytics, Marketing, Operations, and others.

• Are there bundled collections of skills? The page indicates curated bundles (e.g., “Claude Skills”) that group multiple skills together.

• Which platforms are supported? The page explicitly mentions Claude Code, Cursor, Copilot, Windsurf, Zed, and multiple other AI tools (“20+ tools”), plus specific platform groupings shown in the directory.

Alternatives

• Direct plugin/extension marketplaces for each AI tool (vendor marketplaces). These are typically more focused on one platform, whereas agentskill.sh is a cross-platform directory aimed at installing skills across many tools.
• Specialized agent frameworks and registries. Instead of installing pre-built skills, you may define or assemble capabilities within an agent framework, which can be more flexible but requires more setup.
• Manual prompt/tool integration repositories. For teams that prefer full control, repositories that document integrations can be used to wire tools to agents without a centralized “skills” installation workflow.
• General-purpose knowledge management systems with AI plugins. If your goal is primarily managing documents and pages (e.g., Feishu-like knowledge bases), dedicated knowledge management tools may cover parts of the workflow, while agentskill.sh focuses on skill installation for AI agents/tools.