Skip to content
FreeLLMFree AI access, with evidence
FreeLLM / Skills

Discover agent skills with test context

Skills are reusable instructions and workflows, not guaranteed plug-and-play capabilities. We surface evaluation scope and limitations so that a published skill is not mistaken for a fully tested integration.

All modelsProvidersLatest changes

Evidence levels matter

End-to-end test
The workflow ran through a meaningful real task with observable output.
Artifact or task test
Specific behavior was exercised, but does not prove general reliability.
Partial or blocked
A capability was only partly confirmed, or external access prevented completion.

Practical review checklist

  1. Confirm the skill's task and required runtime tools.
  2. Inspect evaluation evidence and failure notes.
  3. Run it on disposable inputs before trusting production documents.
  4. Review generated artifacts and keep sensitive credentials out of prompts.

Continue into the full catalog

Browse tested SkillsFull records and first-party evidence (Chinese UI) Agent workflow recipesCreate a repeatable sequence About our evidence policyWhat counts as a verified claim

English editorial page. Some linked provider records, source quotations and interactive tools remain in their original language; those links are clearly separate from this English route.