skill-recommender

Skill discovery works by automatically scanning your installed Skills and comparing your request against their keywords and descriptions. When you describe a task, the skill-recommender matches your language to relevant Skills and surfaces up to three recommendations without requiring manual searching or prior knowledge of what's available.

Context analysis examines the specific problem you're solving and your current work to identify which installed Skills apply. Rather than generic suggestions, it evaluates skill descriptions and activation keywords against your exact request, ranking matches by relevance to guide you toward the most applicable automation.

Yes. Intelligent recommendations work across all user requests by continuously scanning for skill keywords and comparing installed skill descriptions. The Skill applies universally without requiring special syntax or setup, making automation discovery passive and non-intrusive as you work.

Keyword matching compares your request text against installed skill frontmatter parsed from ~/.claude/skills. The Skill identifies relevant Skills by analyzing descriptions and activation keywords, then returns ranked recommendations. This proposal-only approach prevents unwanted automation while surfacing options for manual activation.

Manual selection requires knowing a Skill exists and explicitly invoking it; skill discovery detects applicability automatically. Intelligent recommendations identify and rank matching Skills from your inventory based on context, eliminating the need to remember or search for the right automation before you can use it.