agent-detector

Task routing detects agent type, complexity level, and appropriate model for each incoming message in real time. The skill categorizes requests as Quick, Standard, or Deep, identifies relevant agents through multi-layer scoring of technology mentions and intent patterns, then outputs a structured Detection Result to guide downstream orchestration.

Multi-layer detection analyzes explicit technology, intent patterns, project context, and file patterns to activate the primary agent and optional supporting agents. This ensures every query reaches the most capable handler and eliminates manual routing overhead.

Yes. Conditional logic intelligently activates QA agents for bug fixes and new implementations while skipping them for pure documentation or design work. The skill applies sophisticated scoring to determine which agents are needed for each request type.

The skill applies layered detection across technology mentions, intent patterns, and project context to score and classify incoming messages. It automatically maps detected complexity levels and identified agents to appropriate models, producing a structured result for task orchestration.

Agent detection extracts technology identifiers, intent signals, project context clues, and file patterns from user input. It uses this layered scoring to determine primary and secondary agents and select the matching model, outputting all results as a Detection Result for downstream processing.