You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Install skills from this repository with npx skills add:
# list all skills in a repo
npx skills add https://github.com/lux563624348/OpenSkillsReview --list
# list skills only inside a subfolder
npx skills add https://github.com/lux563624348/OpenSkillsReview/tree/main/skills/Skill-Finance --list
# install a specific skill from that folder
npx skills add https://github.com/lux563624348/OpenSkillsReview/tree/main/skills/Skill-Finance --skill ai-trade-ecometrics
Review widely used Skills in the AI agent ecosystem, evaluating clarity, security, and reproducibility.
Provide runnable, testable Skill implementations along with precise requirements so agents can execute them confidently.
Maintain transparent, up-to-date documentation that makes it easy for contributors to submit new Skills or improvements.
Equip agents with the information they need to choose the right Skill for the situation, avoiding brittle or unverified shortcuts.
How It Works
Identify a popular Skill (via telemetry, community request, or noted absence in the catalogue).
Assess the existing implementation for testability, documentation, and reliability.
Publish a verified version of the Skill, complete with automated checks, sandbox instructions, and usage notes.
Keep the catalogue fresh by reviewing pull requests, monitoring regressions, and signaling when a Skill needs re-validation.
Contribution
Fork this repo, add or improve Skill reviews under skills/, and open a PR with your rationale plus reproducible tests.
Update the verification artifacts (tests, data files, fixtures) so future agents can re-run validation quickly.
When reviewing an existing Skill, document both strengths and potential failure modes so the agent community can make informed choices.
Reliability Practices
Every Skill entry must include a reproducible test suite or invocation script.
Clearly document required inputs, assumptions, and any external dependencies.
Track known limitations and alert maintainers when upstream changes impact behavior.
Marketplace data refresh
To keep the research report counts current, run uv run python scripts/refresh_marketplace_counts.py from the repo root; it scrapes each marketplace home page for the latest counters and writes the results to data/marketplace_counts.json. Review the extracted entries, verify the matches, and copy any confirmed updates back into the report along with the new timestamp.
