Function useAgentPool

Concurrent agent generation loop as an Effection resource

Runs N agents in parallel using a four-phase tick loop over shared BranchStore infrastructure. Each agent forks from a parent branch, generates tokens, invokes tools, and reports findings.

Four-phase tick loop:

1. PRODUCE — sample all active agents via produceSync() (no async gap)
2. COMMIT — single GPU call via store.commit() for all produced tokens
3. SETTLE — drain settled tool results, batch prefill, reset grammars
4. DISPATCH — execute collected tool calls sequentially via scoped() + call()

Tool dispatch uses scoped() + call() — each tool executes to completion before the next tick, ensuring exclusive llama_context access (no concurrent decode).

Resource semantics: provide() suspends after all agents complete, keeping branches alive so the caller can fork from them (e.g. for verification). Branches are pruned when the scope exits — each branch's ensure() from setupAgent handles cleanup automatically.

For automatic branch cleanup on return, use runAgents instead.