feat(subagent): add configurable subagent model via config.yaml

Allow users to configure a dedicated model for subagents spawned by
delegate_task, so narrowly-scoped subtasks can use a cheaper/faster
model while the parent agent runs on a more powerful one.

Config:
  subagent:
    model: google/gemini-3-flash-preview

Precedence: explicit model arg > config.subagent.model > parent model.

Cherry-picked from PR #751 by Bartok9, rebased onto current main
with conflict resolution and simplified to model-only override
(provider/base_url/api_key stay inherited from parent — covers the
common case of same-provider model swap via OpenRouter).

Closes #609

Co-authored-by: Bartok Moltbot <bartokmoltbot@users.noreply.github.com>

hermes_cli/config.py

@@ -119,6 +119,13 @@ DEFAULT_CONFIG = {
         },
     },
     
+    # Subagent configuration — model/provider for tasks spawned via delegate_task.
+    # By default subagents inherit the parent agent's model and provider.
+    # Set "model" to use a cheaper/faster model for delegated subtasks.
+    "subagent": {
+        # "model": "google/gemini-3-flash-preview",
+    },
+
     "display": {
         "compact": False,
         "personality": "kawaii",

tools/delegate_tool.py

@@ -78,6 +78,15 @@ def _strip_blocked_tools(toolsets: List[str]) -> List[str]:
     return [t for t in toolsets if t not in blocked_toolset_names]
 
 
+def _get_subagent_config() -> Dict[str, Any]:
+    """Load subagent config from CLI_CONFIG if available."""
+    try:
+        from cli import CLI_CONFIG
+        return CLI_CONFIG.get("subagent", {})
+    except Exception:
+        return {}
+
+
 def _build_child_progress_callback(task_index: int, parent_agent, task_count: int = 1) -> Optional[callable]:
     """Build a callback that relays child agent tool calls to the parent display.
 
@@ -199,10 +208,15 @@ def _run_single_child(
         # count toward the session-wide limit.
         shared_budget = getattr(parent_agent, "iteration_budget", None)
 
+        # Subagent model override from config.
+        # Precedence: explicit model arg > config.subagent.model > parent model
+        subagent_cfg = _get_subagent_config()
+        effective_model = model or subagent_cfg.get("model") or parent_agent.model
+
         child = AIAgent(
             base_url=parent_agent.base_url,
             api_key=parent_api_key,
-            model=model or parent_agent.model,
+            model=effective_model,
             provider=getattr(parent_agent, "provider", None),
             api_mode=getattr(parent_agent, "api_mode", None),
             max_iterations=max_iterations,