Guides teams through creating and refining formal SKILL.md documents.
npx skills add https://github.com/memect/kc --skill skill-creator-memect
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/memect/kc/tree/main/template/skills/zh/skill-creator Command: npx skills add https://github.com/memect/kc --skill skill-creator-memect
The Skill Creator helps teams draft, version, and improve skills by providing a structured workflow to ideate, write SKILL.md, run evaluation loops, and iterate based on feedback and benchmarks.
Draft an initial SKILL.md for a new capability, scaffold a minimal eval set, and run the end-to-end evaluation and improvement loop to iterate toward a polished skill.
High-intent search queries and answers about installing and using this skill.
To create and iterate on Claude skills, you can use a structured workflow to draft a SKILL.md file, run trigger tests, generate reviewer feedback, and execute evaluation loops to refine the skill based on benchmarks.
The best way to benchmark skill performance is to compare skill variants using automated scoring and informed recommendations, capturing iteration history to push quality upward through continuous evaluation loops.
Yes, you need Python and the PyYAML dependency to draft a formal SKILL.md with frontmatter, as the skill creation and refinement workflow relies on these libraries to scaffold and process evaluation sets.
Yes, you can run end-to-end evaluation loops without writing custom test scripts by using the provided scripts to scaffold a minimal eval set, execute trigger tests, and generate reviewer feedback automatically.
Capturing iteration history improves the skill refinement process by tracking evaluation loops and reviewer feedback, allowing teams to compare skill variants against benchmarks and make informed recommendations for targeted improvements.
amitmishra-bylddAudits prose and outputs actionable fixes in a three-column table.
timurgaleevMonitor post-deploy web apps for defects and performance regressions.
srivtxAutomate code review for local changes and GitHub pull requests.
woodfishhhhRefine LaTeX research paper drafts via simulated peer review with accept/revert rules.
