skill-developer

To create Claude Code skills, you define the skill structure, configure YAML frontmatter, and set up trigger types with enforcement levels in skill-rules.json. This enables the auto-activation system to match patterns and proactively suggest your skill.

UserPromptSubmit and PreToolUse hook mechanisms provide proactive suggestions and guardrails during skill development. They intercept user prompts and tool executions, allowing your Claude Code skills to enforce rules and trigger contextually based on pattern matching.

False positives and non-triggering skills in the auto-activation system stem from misconfigured pattern matching or rule configuration issues. Debugging involves checking your YAML frontmatter, verifying trigger types, and validating your skill-rules.json against Anthropic's best practices.

The 500-line rule and progressive disclosure are Anthropic best practices for skill structure. They ensure your Claude Code skills remain maintainable by limiting file length and revealing detailed instructions only when contextually required by the auto-activation system.

Yes, you can build an API documentation helper skill by defining specific triggers and rules in skill-rules.json. This configures the auto-activation system to match API endpoint queries and proactively suggest your documentation skill.

Debugging rule configuration and pattern matching requires validating your YAML frontmatter and trigger types against Anthropic's best practices. Proper configuration ensures the auto-activation system correctly interprets enforcement levels and avoids false positives.