refactor: allow model/context_window_tokens override in from_config()

- Pop model and context_window_tokens from extra kwargs before
  forwarding to __init__, allowing callers like _run_gateway to
  pass snapshot-derived values instead of config defaults
- _run_gateway now explicitly passes model/context_window_tokens
  from provider_snapshot to preserve pre-refactor behavior

nanobot/agent/loop.py

@@ -360,13 +360,15 @@ class AgentLoop:
             bus = MessageBus()
         defaults = config.agents.defaults
         provider = extra.pop("provider", None) or make_provider(config)
+        model = extra.pop("model", None) or defaults.model
+        context_window_tokens = extra.pop("context_window_tokens", None) or defaults.context_window_tokens
         return cls(
             bus=bus,
             provider=provider,
             workspace=config.workspace_path,
-            model=defaults.model,
+            model=model,
             max_iterations=defaults.max_tool_iterations,
-            context_window_tokens=defaults.context_window_tokens,
+            context_window_tokens=context_window_tokens,
             context_block_limit=defaults.context_block_limit,
             max_tool_result_chars=defaults.max_tool_result_chars,
             provider_retry_mode=defaults.provider_retry_mode,

nanobot/cli/commands.py

@@ -655,6 +655,8 @@ def _run_gateway(
     agent = AgentLoop.from_config(
         config, bus,
         provider=provider_snapshot.provider,
+        model=provider_snapshot.model,
+        context_window_tokens=provider_snapshot.context_window_tokens,
         cron_service=cron,
         session_manager=session_manager,
         image_generation_provider_configs={