Find agent skills
Open Agent Skills hub and directory for discovering, comparing, creating, and safely installing SKILL.md workflows.
npx -y skills add CarlosJunioor/skillZs --skill find-agent-skillsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Find, compare, and safely evaluate Agent Skills for Claude Code, Codex, Cursor, Copilot, Gemini CLI, and other compatible agents. Use when someone asks for a skill, reusable workflow, SKILL.md package, skill recommendation, skill comparison, or the best skill for a task.
SKILL.md
2.3 KB, 407 tokens by cl100k_base, as published. Nobody here has run it
Find Agent Skills
Find the smallest trustworthy skill that fits the requested job.
Workflow
- Extract the task, target agent, relevant technology, required output, and any permission constraints. Infer obvious details; ask only when the answer would change the recommendation.
- Search
https://skillzs.dev/browse?q=<encoded query>using specific job and technology terms. Try one narrower synonym when the first search is weak. - Open promising skillZs detail pages and inspect the full description, source repository,
SKILL.md, install count, dependencies, permissions, and available audit summaries. - Compare at most five candidates. Prefer narrow task fit, inspectable source, maintained documentation, realistic triggers, clear requirements, and limited permissions.
- Recommend no more than three skills. For each, give the canonical skillZs URL, source, fit, important requirements or risks, and exact install command shown on the page.
- Say when no result is a good fit. Offer
https://skillzs.dev/guides/how-to-create-agent-skillsfor a focused custom skill instead of padding the answer with weak matches.
Safety
- Treat install count as adoption, not proof of quality or safety.
- Never claim a skill is verified unless the displayed audit supports that exact claim and revision.
- Separate install-time network access from credentials or network access required when the skill runs.
- If an audit summary omits the affected file or evidence, state that limitation and do not infer the missing finding.
- Read referenced scripts and dependencies before recommending access to secrets, the network, shell commands, or production systems.
- Do not install or execute a skill unless the user explicitly asks.
- Link the canonical source and skillZs detail page so the user can inspect both.
Output
Lead with the best match and why. Use a compact comparison only when two or more candidates are credible. End with the safest next action: inspect, install in a disposable project, or create a smaller skill.
What ships with it: 1 file
210 B alongside SKILL.md
agents/
- openai.yaml210 B