[One sentence: what this skill does — Claude uses this to decide when to invoke it]
Placement: .claude/skills/[kebab-case-name]/SKILL.md
Naming: kebab-case, ≤64 chars, lowercase letters/numbers/hyphens only
Target: Under 200 lines after placeholders are filled
Rule: `description` is required — Claude uses it to decide when to invoke
Rule: `user-invocable: false` is the default for migrated skills (auto-invoked by Claude)
Rule: Every instruction must be specific and verifiable
Source attribution: add <!-- Migrated from [source-file]:lines [N-M] --> below frontmatter
-->
<!-- Migrated from [source-file]:lines [N-M] --> <!-- CONDITIONAL: Use `disable-model-invocation: true` instead of `user-invocable: false` ONLY when the skill is heavy, rare, or has side effects (invoked <20% of sessions). With this flag: zero passive context cost; user must invoke manually via slash command. Replace `user-invocable: false` in frontmatter with: disable-model-invocation: true --> <!-- CONDITIONAL: Add `context: fork` to frontmatter ONLY when the skill performs context-heavy isolated analysis (similar to a subagent task). With this flag: skill runs in a forked context; parent context cost is zero. Add to frontmatter: context: fork --> <!-- CONDITIONAL: Add `allowed-tools` to frontmatter ONLY when the skill requires specific tools restricted to its scope. Add to frontmatter: allowed-tools: [Read, Glob, Grep] --> <!-- CONDITIONAL: Include additional sections ONLY if the migrated content has distinct phases or steps that benefit from headers. -->[One-two sentences describing what this skill does and when it applies. Only include context that helps the agent understand why this skill exists.]