Agent skill resolver

Describe the task. Get the right reusable skill.

OpenAgentSkill turns vague agent work into a concrete install plan: recommended skill, alternatives, Trust Score, audit notes, safety policy, and target-specific handoff.

Best skill

Score

Agent Proven

Receipt

Install handoff

Resolved decision

Last30days Skill

Research recent cross-source changes across Reddit, X, YouTube, Hacker News, and the web.

99/100 task fitNeeds first agent outcome72/100 Trust88/100 audit

Install

npx skills add mvanhorn/last30days-skill

Safety gate

Reviewed
Review the audit page, then allow agent install in a sandboxed workflow.

Open skill Audit page Text API Install receipt

Task fit

Matches the job your agent needs to do

Searches scenario, tags, repository metadata, category, and use-case signals before choosing a skill.

Trust

Ranks by quality, adoption, and maintenance

Combines GitHub stars, freshness, license clarity, install readiness, and OpenAgentSkill Trust Score.

Agent Proven

Uses real run outcomes when available

Surfaces reported success, recent failures, install success, output quality, and production-use signals.

Install

Returns agent-ready handoffs

Provides CLI command, target agent prompt, detail page, audit page, and safer next steps.

Agent Resolve

Describe the task. Get one skill plan.

API ready

Try one

Recommendation

Resolve output will appear here

The registry returns a decision, not just links.

Your agent gets one recommended skill, alternatives, install command, Trust Score, audit summary, and a safety policy before it acts.

Plain text API