Automate Claude Code skill creation from SKILL.md authoring to evaluation.
npx skills add https://github.com/majiayu000/spellbook --skill skill-creator-majiayu000
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/majiayu000/spellbook/tree/main/skills/skill-creator Command: npx skills add https://github.com/majiayu000/spellbook --skill skill-creator-majiayu000
The Skill helps teams rapidly author Claude Code skills, organize their frontmatter and body, and drive iterative improvement through automated evaluation loops so that skills become more reliable and repeatable.
Install the skill and run the built-in evaluation loop to iteratively improve its triggering description.
High-intent search queries and answers about installing and using this skill.
You create a Claude Code skill by drafting a SKILL.md file with structured frontmatter and a concise body. This skill automates that end-to-end authoring process while supporting progressive disclosure and optional bundled resources like scripts and references.
Automated evaluation runs iterative feedback loops on your SKILL.md to optimize its triggering description. By benchmarking performance and integrating feedback, it refines the skill to become more reliable and repeatable across different runtimes.
Progressive disclosure in SKILL.md authoring structures information using frontmatter and a concise body, deferring complex logic to optional bundled resources. This keeps the skill lightweight while still powering advanced, repeatable workflows.
Yes, you can bundle optional resources such as scripts, references, and assets with your skill. These components support advanced workflows and are packaged alongside the SKILL.md for distribution and reuse across runtimes.
You need the anthropic and PyYAML dependencies to utilize the automated evaluation loops, frontmatter parsing, and description optimization features that drive iterative skill refinement and benchmarking.
The best way to optimize a triggering description is running a built-in automated evaluation loop. This benchmarks the skill, collects feedback, and iteratively refines the description to ensure reliable and repeatable task triggering.
TheWayWithinDesign and automate multi-step SaaS user onboarding flows with activation milestones.
hasnaxyzCreate and package Claude skill directories with valid SKILL.md frontmatter.
lucasfthDesign, validate, and iteratively improve Claude Skills with SKILL.md and eval loops.
mwolff328-stackDevelop, test, and refine custom AI skills through iterative evaluation loops.
