domain-classifier

Automatic task routing classifies a task description into multiple domains using a 12-domain taxonomy, assigning individual confidence scores to prioritize and trigger the appropriate skills and subagents.

Yes, multi-domain classification detects overlapping areas such as UI/UX, API/Backend, and Database simultaneously, providing multi-label results with individual confidence scores for each detected domain.

Confidence-based routing flags ambiguous inputs by evaluating domain confidence scores, suggesting clarifying questions to resolve uncertainty before recommending specific skills, subagents, or workflows.

Domain classification supports multi-repo tasks by orchestrating phased workflows, enabling contract validation across repositories while applying its 12-domain taxonomy to route tasks accurately.

Classifying a task like adding a login page with OAuth support involves evaluating it against multiple domains, yielding a primary domain, individual confidence scores, and suggested routing to relevant subagents.

Domain routing handles both auto-trigger and manual-trigger inputs by evaluating the task description against its taxonomy, ensuring accurate skill prioritization regardless of how the workflow is initiated.