improve-skill

You validate agent skill markdown files by running them through an iterative sandbox implementation that executes automated checks for TypeScript, linting, building, and browser tests to catch code and documentation errors.

Refining skill documentation and code snippets is done by implementing the skill in a sandboxed environment, reflecting on errors from automated validation checks, and updating the markdown until all acceptance criteria are met.

Yes, you can test skill setup instructions in a sandbox by running the implementation loop which automatically executes validation checks and reflects on errors to ensure the instructions produce error-free code.

Automated skill validation works with TypeScript and linting checks by running comprehensive validation tests within a sandboxed environment to identify missing steps or bugs in the markdown.

Agent skill markdown fails during implementation when it contains issues in documentation, code snippets, or setup instructions that cause errors during automated TypeScript, linting, building, or browser validation checks.

Iterative markdown refinement for skill development requires a sandboxed environment to run validation checks, meaning it depends on the sandbox supporting the specific testing frameworks and build tools needed by the skill.