Automate drafting, evaluating, and iterating Claude Skills with SKILL.md.
npx skills add https://github.com/cagdotin/agents --skill skill-creator-cagdotin
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/cagdotin/agents/tree/main/.agents/skills/skill-creator Command: npx skills add https://github.com/cagdotin/agents --skill skill-creator-cagdotin
Skill creators help teams quickly draft, test, and refine Claude Skills through an end-to-end loop of drafting SKILL.md, running trigger evaluations, and optimizing descriptions based on feedback and benchmarks.
Draft a new skill, then run an initial evaluation cycle to see triggering behavior and iterate.
High-intent search queries and answers about installing and using this skill.
You can automate Claude Code skill creation by running an end-to-end loop that drafts SKILL.md files, executes trigger evaluations, and optimizes descriptions iteratively. It provides tooling for prompt evaluation, results analysis, and description optimization across cycles.
An eval loop for skill benchmarking is an iterative cycle of testing trigger behavior, analyzing grading data, and refining prompts. It supports train/test evaluation, history tracking, and produces timing and metrics artifacts for governance.
To run skill evaluation workflows, you need the pyyaml dependency installed and the provided scripts configured. These scripts generate the scaffolding for outputs, timing, and grading data required to execute the benchmarking cycles.
Yes, you can track and compare skill descriptions across iterations. The workflow supports history tracking and produces monitoring artifacts like evals, timing, and metrics to benchmark and optimize description performance over iterative cycles.
The best way to refine Claude Code trigger behavior is through an automated evaluation loop that drafts, tests, and iterates descriptions based on feedback. This approach uses benchmarking and grading data to systematically optimize triggering accuracy.
Monitoring artifacts are needed for skill benchmarking to provide governance data and track performance over time. They capture evaluation results, timing, and grading metrics, ensuring you can accurately analyze and optimize trigger behavior across iterations.
Fudan-SMI-labPreinstall PyYAML before torch-npu to prevent transitive dependency failures.
dy9759Draft, evaluate, and benchmark SKILL.md-driven workflows with bundled scripts.
NeerajCodzCreates, modifies, and optimizes Claude skills with benchmarking.
codeboltaiCreate and refine Claude skills with structured SKILL.md templates and packaging workflows.
