refactor(agent): capture subagent runtime before spawn

nanobot/agent/loop.py

@@ -340,11 +340,8 @@ class AgentLoop:
         self._file_state_store = FileStateStore()
         self.runner = AgentRunner()
         self.subagents = SubagentManager(
-            provider=provider,
             workspace=workspace,
             bus=bus,
-            model=self.model,
-            context_window_tokens=self.context_window_tokens,
             tools_config=_tc,
             max_tool_result_chars=self.max_tool_result_chars,
             restrict_to_workspace=restrict_to_workspace,
@@ -495,7 +492,6 @@ class AgentLoop:
         self.provider = provider
         self.model = model
         self.context_window_tokens = context_window_tokens
-        self.subagents.set_provider(provider, model, context_window_tokens)
         self.consolidator.set_provider(provider, model, context_window_tokens)
         self._sync_replay_max_messages()
         self._provider_signature = snapshot.signature

nanobot/agent/subagent.py

@@ -12,14 +12,18 @@ from loguru import logger
 
 from nanobot.agent.hook import AgentHook, AgentHookContext
 from nanobot.agent.runner import AgentRunner, AgentRunSpec
-from nanobot.agent.tools.context import ToolContext
+from nanobot.agent.tools.context import (
+    RequestContext,
+    ToolContext,
+    bind_request_context,
+    reset_request_context,
+)
 from nanobot.agent.tools.file_state import FileStates
 from nanobot.agent.tools.loader import ToolLoader
 from nanobot.agent.tools.registry import ToolRegistry
 from nanobot.bus.events import InboundMessage
 from nanobot.bus.queue import MessageBus
 from nanobot.config.schema import AgentDefaults, ToolsConfig
-from nanobot.providers.base import LLMProvider
 from nanobot.security.workspace_access import (
     WorkspaceScope,
     bind_workspace_scope,
@@ -77,12 +81,9 @@ class SubagentManager:
 
     def __init__(
         self,
-        provider: LLMProvider,
         workspace: Path,
         bus: MessageBus,
         max_tool_result_chars: int,
-        model: str | None = None,
-        context_window_tokens: int | None = None,
         tools_config: ToolsConfig | None = None,
         restrict_to_workspace: bool = False,
         disabled_skills: list[str] | None = None,
@@ -92,11 +93,8 @@ class SubagentManager:
         llm_wall_timeout_for_session: Callable[[str | None], float | None] | None = None,
     ):
         defaults = AgentDefaults()
-        self.provider = provider
         self.workspace = workspace
         self.bus = bus
-        self.model = model or provider.get_default_model()
-        self.context_window_tokens = context_window_tokens or defaults.context_window_tokens
         self.tools_config = tools_config or ToolsConfig()
         self.max_tool_result_chars = max_tool_result_chars
         self.restrict_to_workspace = restrict_to_workspace
@@ -152,17 +150,6 @@ class SubagentManager:
         ToolLoader().load(ctx, registry, scope="subagent")
         return registry
 
-    def set_provider(
-        self,
-        provider: LLMProvider,
-        model: str,
-        context_window_tokens: int | None = None,
-    ) -> None:
-        self.provider = provider
-        self.model = model
-        if context_window_tokens is not None:
-            self.context_window_tokens = context_window_tokens
-
     async def spawn(
         self,
         task: str,
@@ -173,8 +160,12 @@ class SubagentManager:
         origin_message_id: str | None = None,
         temperature: float | None = None,
         workspace_scope: WorkspaceScope | None = None,
+        *,
+        runtime: LLMRuntime,
     ) -> str:
         """Spawn a subagent to execute a task in the background."""
+        if temperature is not None:
+            runtime = runtime.with_generation_overrides(temperature=temperature)
         task_id = str(uuid.uuid4())[:8]
         display_label = label or task[:30] + ("..." if len(task) > 30 else "")
         origin = {"channel": origin_channel, "chat_id": origin_chat_id, "session_key": session_key}
@@ -194,8 +185,8 @@ class SubagentManager:
                 display_label,
                 origin,
                 status,
+                runtime,
                 origin_message_id,
-                temperature,
                 workspace_scope,
             )
         )
@@ -223,8 +214,8 @@ class SubagentManager:
         label: str,
         origin: dict[str, str],
         status: SubagentStatus,
+        runtime: LLMRuntime,
         origin_message_id: str | None = None,
-        temperature: float | None = None,
         workspace_scope: WorkspaceScope | None = None,
     ) -> None:
         """Execute the subagent task and announce the result."""
@@ -253,13 +244,15 @@ class SubagentManager:
                 if self._llm_wall_timeout_for_session
                 else None
             )
+            request_token = bind_request_context(RequestContext(
+                channel=origin["channel"],
+                chat_id=origin["chat_id"],
+                message_id=origin_message_id,
+                session_key=sess_key,
+                runtime=runtime,
+            ))
             token = bind_workspace_scope(workspace_scope) if workspace_scope is not None else None
             try:
-                runtime = LLMRuntime.capture(
-                    self.provider,
-                    self.model,
-                    context_window_tokens=self.context_window_tokens,
-                ).with_generation_overrides(temperature=temperature)
                 result = await self.runner.run(AgentRunSpec(
                     initial_messages=messages,
                     tools=tools,
@@ -279,6 +272,7 @@ class SubagentManager:
             finally:
                 if token is not None:
                     reset_workspace_scope(token)
+                reset_request_context(request_token)
             status.phase = "done"
             status.stop_reason = result.stop_reason
 

nanobot/agent/tools/spawn.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 from typing import TYPE_CHECKING, Any
 
-from nanobot.agent.tools.base import Tool, tool_parameters
+from nanobot.agent.tools.base import Tool, ToolResult, tool_parameters
 from nanobot.agent.tools.context import current_request_context
 from nanobot.agent.tools.schema import NumberSchema, StringSchema, tool_parameters_schema
 from nanobot.security.workspace_access import current_workspace_scope
@@ -70,20 +70,19 @@ class SpawnTool(Tool):
                 f"to complete before spawning a new one."
             )
         request_ctx = current_request_context()
-        origin_channel = request_ctx.channel if request_ctx is not None else "cli"
-        origin_chat_id = request_ctx.chat_id if request_ctx is not None else "direct"
-        session_key = (
-            request_ctx.session_key or f"{origin_channel}:{origin_chat_id}"
-            if request_ctx is not None
-            else "cli:direct"
-        )
+        if request_ctx is None or request_ctx.runtime is None:
+            return ToolResult.error("Error: spawn requires an active model runtime")
+        origin_channel = request_ctx.channel
+        origin_chat_id = request_ctx.chat_id
+        session_key = request_ctx.session_key or f"{origin_channel}:{origin_chat_id}"
         return await self._manager.spawn(
             task=task,
+            runtime=request_ctx.runtime,
             label=label,
             origin_channel=origin_channel,
             origin_chat_id=origin_chat_id,
             session_key=session_key,
-            origin_message_id=request_ctx.message_id if request_ctx is not None else None,
+            origin_message_id=request_ctx.message_id,
             temperature=temperature,
             workspace_scope=current_workspace_scope(),
         )