Krisp Voice TranslationSpeech-to-speech translation API
Promoted
Maker
📌
Hey Product Hunt,
I built OpenAgentSkill because the AI agent ecosystem is getting harder to navigate. There are useful skills scattered across GitHub, Claude Code workflows, Codex workflows, browser automation tools, RAG tools, data tools, and productivity repos. But before an agent installs something, stars alone do not answer the important questions:
- Is this skill relevant to the task?
- Is it maintained?
- Is it safe enough to use?
- What are the tradeoffs?
- Which agent workflow does it fit?
OpenAgentSkill indexes high-signal skill repositories, excludes MCP repos from automated imports, and adds quality, trust, audit, trending, hot, and recommendation layers on top.
What is live today:
- 980+ indexed AI agent skills
- Skill search and ranking pages
- Audit reports with risk and trust signals
- Trending and hot rankings from daily activity
- README badges for skill authors
- Agent-friendly APIs for search, ranking, audit, and task-based recommendation
Example:
GET /api/agent/recommend?task=scrape+websites+and+extract+tables
I would love feedback from builders here:
1. What trust signals should matter most before an agent installs a skill?
2. Which agent surfaces should we support more deeply: Claude Code, Codex, Cursor, Windsurf, or something else?
3. Would you use a public audit badge in your own skill README?
Thanks for checking it out.
