claude-code-skill-expert

Create a Claude Code skill with structured frontmatter containing name and description, progressive disclosure in the workflow, precise output templates, reference separation, and robust error handling. Start by defining clear metadata, then build skill logic incrementally to meet Anthropic's standards for token efficiency and context usage.

A SKILL.md file requires frontmatter with skill name and description, clear workflow steps with progressive loading, actionable metadata, validation checkpoints for examples and templates, and error handling guidance. This structure ensures scalability and helps users understand the skill's purpose and usage.

Audit a Claude Code skill by checking frontmatter completeness, verifying progressive disclosure in workflow design, reviewing output templates for precision, confirming reference separation, assessing token efficiency, and validating error handling. This process identifies gaps and ensures alignment with Anthropic's quality standards.

Progressive disclosure in skill design prevents cognitive overload by revealing information and options incrementally as users advance through workflows. This approach improves usability, reduces token consumption, and keeps skills focused and maintainable while supporting both novice and advanced users.

Common issues include trigger misconfiguration, inefficient context usage, incomplete frontmatter, and weak error handling. Troubleshoot by validating trigger logic against use cases, profiling token consumption, auditing metadata completeness, and ensuring error messages guide users toward recovery.