router

Task routing analyzes your request's intent and project context to automatically direct it to the most appropriate tool. It extracts action and domain from your input, checks real-time project state, scores confidence in the best match, and executes the selection—eliminating manual tool selection and ensuring efficient workflows.

Yes. Routing manages sequential and parallel tool invocation for multi-step workflows. When you request 'fix my code and commit it', it detects TypeScript errors, runs the fix tool, then executes git commit with conventional messaging—automating what would otherwise require manual steps across separate tools.

The router combines intent analysis—extracting action, domain, scope, and urgency from your request—with context awareness of real-time project state including Git status, diagnostics, and file types. A decision engine applies pattern matching and confidence scoring to select the optimal tool and communicate its reasoning transparently.

Yes. Routing applies across feature development, bug fixes, code reviews, testing, documentation, and deployment. It handles diverse intents and project states through context-aware decision logic, so the same routing mechanism works consistently whether you're fixing issues, building features, or preparing releases.

The router uses confidence scoring within its decision engine to assess each potential match. It communicates routing rationale transparently, allowing you to understand why a particular tool was selected and providing visibility into edge cases where confidence may be lower than typical routing scenarios.