skill-developer

Create a skill by setting up a SKILL.md file in .claude/skills/{skill-name}/, then add its configuration to .claude/skills/skill-rules.json. Test activation triggers using echo '{"prompt":"your test prompt"}' | npx tsx .claude/hooks/skill-activation-prompt.ts to verify it responds correctly.

Trigger patterns define how skills activate—through keywords, intent, file paths, or content matching. Enforcement levels control whether a skill runs always, conditionally, or only on demand. Both are configured in your skill's frontmatter and rule definitions to govern when and how your skill executes.

Skill activation failures usually stem from misconfigured triggers, incorrect enforcement levels, or pattern mismatches. Diagnose by testing your prompt directly with the skill-activation-prompt.ts hook, checking your trigger definitions in skill-rules.json, and verifying your SKILL.md frontmatter syntax.

Yes. Hooks enable you to intercept and control skill execution at specific points, while session tracking maintains state across invocations. Both are built into the skill governance framework to support robust testing, debugging, and consistent behavior across documentation and references.

You need npx, tsx, and jq. These tools handle skill execution, TypeScript transpilation, and JSON configuration parsing respectively. All three are required for creating, testing, and validating skill structure and trigger patterns.

Define guardrail and domain skills that encode your team's patterns and standards. Configure them with content and file pattern triggers, set appropriate enforcement levels, and test them within your skill framework to ensure consistent AI-assisted development practices across your project.