About this skill

Problem Addressed

In WorkBuddy, choosing among many similarly named experts can be hard. This skill shifts expert selection from keyword lookup to task-capability matching. It identifies which single expert or expert team covers the user's task and provides an actionable summoning path.

How It Works

The skill first confirms the task description and asks for clarification when it is vague. It then reads the local compact index expert-compact.txt and performs semantic filtering across 474 experts and 14 industry categories, prioritizing:
- Capability fit: whether the expert's role and description cover the required work;
- Task complexity: multi-step or multi-role workflows point to expert teams, while single clear requests point to individual experts;
- Install state: installed experts are preferred when capabilities are comparable;
- False-match avoidance: matching is based on professional capability, not surface keywords.

It then reads candidate details from expert-index.json, including description, defaultInitPrompt, and team members. The final answer gives 2–4 recommendations, explains why each is suitable, shows how to summon it, and, for teams, describes member roles and handoff workflow. A one-sentence final choice suggestion follows.

Boundaries

The skill relies on the WorkBuddy expert library and local indexes. It does not create experts or bypass expert-center install state. If no close match exists, it says so and suggests browsing the expert center or creating a custom expert; the index may need refreshing after library updates.