What problem does it solve? It prevents wasteful or incorrect subagent delegation by enforcing strict gates before any spawn, keeping work inline when coordination overhead exceeds the benefit. ## Core Features & Use Cases - Dispatch Rationales: Spawn only for parallel time wins (≥3 independent workstreams), verifiable context offloading, or explicit user request. - Pre-Spawn Checklist: A mandatory five-point checklist covering rationale, parallelism, offload size, profile, and foreground/background mode before every dispatch. - Model Routing: Maps task shapes (scout, doc lookup, implementation, synthesis, test runs) to fast, mid, or strong model tiers. - Use Case: Before launching a large codebase exploration, the agent checks the offloading exception gate (more than 5 file reads across more than 2 modules) and dispatches a read-only scout that returns a distilled brief instead of flooding the parent context. ## Quick Start Before spawning any subagent or running a large codebase scan, apply the subagents dispatch gates and pre-spawn checklist to decide whether to delegate or stay inline.
