Bug

_dispatch_delegate_task() drops the model parameter when forwarding the LLM's tool call to delegate_task(). The schema advertises model as a field and the registry accepts it, but the dispatch layer silently discards it.

This affects ALL providers — any user setting a per-invocation model override (or relying on delegation.model config) sees the child spawn as the parent's model instead.

Root Cause

The DELEGATE_TASK_SCHEMA includes:

"model": {"type": "object", "description": "Per-invocation model override..."}

The LLM correctly includes model={"model": "deepseek-v4-flash"} in its tool call args. But _dispatch_delegate_task() forwards goal, context, toolsets, tasks, max_iterations, acp_command, acp_args, role, parent_agent — everything except model.

The rest of the pipeline is correct (config → _resolve_delegation_credentials → _build_child_agent → AIAgent.__init__), but the LLM tool call path bypasses it entirely.

Fix

One line: add model=function_args.get("model") to the _delegate_task() call.

Verification

• Parent model: deepseek-v4-pro
• Config: delegation.model: deepseek-v4-flash
• Child spawn via delegate_task(goal="Report your model"):
  • Before fix: child runs as deepseek-v4-pro (inherits parent)
  • After fix: child runs as deepseek-v4-flash (config respected) ✓

Confirmed via live debug tracing across config read, credential resolution, AIAgent construction, and run_conversation. Every checkpoint in the pipeline produces the correct value with this fix.