glaip-sdk 0.0.20__py3-none-any.whl → 0.7.7__py3-none-any.whl

glaip_sdk/runner/langgraph.py

@@ -0,0 +1,872 @@
+"""LangGraph-based runner for local agent execution.
+
+This module provides the LangGraphRunner which executes glaip-sdk agents
+locally via the aip-agents LangGraphReactAgent, without requiring the AIP server.
+
+Authors:
+    Christian Trisno Sen Long Chen (christian.t.s.l.chen@gdplabs.id)
+
+Example:
+    >>> from glaip_sdk.runner import LangGraphRunner
+    >>> from glaip_sdk.agents import Agent
+    >>>
+    >>> runner = LangGraphRunner()
+    >>> agent = Agent(name="my-agent", instruction="You are helpful.")
+    >>> result = runner.run(agent, "Hello, world!")
+"""
+
+from __future__ import annotations
+
+import asyncio
+import inspect
+import logging
+from dataclasses import dataclass
+from typing import TYPE_CHECKING, Any
+
+from gllm_core.utils import LoggerManager
+
+from glaip_sdk.client.run_rendering import AgentRunRenderingManager
+from glaip_sdk.hitl import PauseResumeCallback
+from glaip_sdk.runner.base import BaseRunner
+from glaip_sdk.runner.deps import (
+    check_local_runtime_available,
+    get_local_runtime_missing_message,
+)
+from glaip_sdk.utils.tool_storage_provider import build_tool_output_manager
+
+if TYPE_CHECKING:
+    from langchain_core.messages import BaseMessage
+
+    from glaip_sdk.agents.base import Agent
+
+
+_AIP_LOGS_SWALLOWED = False
+
+
+def _swallow_aip_logs(level: int = logging.ERROR) -> None:
+    """Consume noisy AIPAgents logs once (opt-in via runner flag)."""
+    global _AIP_LOGS_SWALLOWED
+    if _AIP_LOGS_SWALLOWED:
+        return
+    prefixes = ("aip_agents.",)
+
+    def _silence(name: str) -> None:
+        lg = logging.getLogger(name)
+        lg.handlers = [logging.NullHandler()]
+        lg.propagate = False
+        lg.setLevel(level)
+
+    # Silence any already-registered loggers under the given prefixes
+    for logger_name in logging.root.manager.loggerDict:
+        if any(logger_name.startswith(prefix) for prefix in prefixes):
+            _silence(logger_name)
+
+    # Also set the base prefix loggers so future children inherit silence
+    for prefix in prefixes:
+        _silence(prefix.rstrip("."))
+    _AIP_LOGS_SWALLOWED = True
+
+
+logger = LoggerManager().get_logger(__name__)
+
+
+def _convert_chat_history_to_messages(
+    chat_history: list[dict[str, str]] | None,
+) -> list[BaseMessage]:
+    """Convert chat history dicts to LangChain messages.
+
+    Args:
+        chat_history: List of dicts with "role" and "content" keys.
+            Supported roles: "user"/"human", "assistant"/"ai", "system".
+
+    Returns:
+        List of LangChain BaseMessage instances.
+    """
+    if not chat_history:
+        return []
+
+    from langchain_core.messages import AIMessage, HumanMessage, SystemMessage  # noqa: PLC0415
+
+    messages: list[BaseMessage] = []
+    for msg in chat_history:
+        role = msg.get("role", "").lower()
+        content = msg.get("content", "")
+
+        if role in ("user", "human"):
+            messages.append(HumanMessage(content=content))
+        elif role in ("assistant", "ai"):
+            messages.append(AIMessage(content=content))
+        elif role == "system":
+            messages.append(SystemMessage(content=content))
+        else:
+            # Default to human message for unknown roles
+            logger.warning("Unknown chat history role '%s', treating as user message", role)
+            messages.append(HumanMessage(content=content))
+
+    return messages
+
+
+@dataclass(frozen=True, slots=True)
+class LangGraphRunner(BaseRunner):
+    """Runner implementation using aip-agents LangGraphReactAgent.
+
+    Current behavior:
+    - Execute via `LangGraphReactAgent.arun_sse_stream()` (normalized SSE-compatible stream)
+    - Route all events through `AgentRunRenderingManager.async_process_stream_events`
+      for unified rendering between local and remote agents
+
+    Attributes:
+        default_model: Model name to use when agent.model is not set.
+            Defaults to "gpt-4o-mini".
+    """
+
+    default_model: str = "openai/gpt-4o-mini"
+
+    def run(
+        self,
+        agent: Agent,
+        message: str,
+        verbose: bool = False,
+        runtime_config: dict[str, Any] | None = None,
+        chat_history: list[dict[str, str]] | None = None,
+        *,
+        swallow_aip_logs: bool = True,
+        **kwargs: Any,
+    ) -> str:
+        """Execute agent synchronously and return final response text.
+
+        Args:
+            agent: The glaip_sdk Agent to execute.
+            message: The user message to send to the agent.
+            verbose: If True, emit debug trace output during execution.
+                Defaults to False.
+            runtime_config: Optional runtime configuration for tools, MCPs, etc.
+                Defaults to None. (Implemented in PR-04+)
+            chat_history: Optional list of prior conversation messages.
+                Each message is a dict with "role" and "content" keys.
+                Defaults to None.
+            swallow_aip_logs: When True (default), silence noisy logs from aip-agents,
+                gllm_inference, OpenAILMInvoker, and httpx. Set to False to honor user
+                logging configuration.
+            **kwargs: Additional keyword arguments passed to the backend.
+
+        Returns:
+            The final response text from the agent.
+
+        Raises:
+            RuntimeError: If the local runtime dependencies are not available.
+            RuntimeError: If no final response is received from the agent.
+        """
+        if not check_local_runtime_available():
+            raise RuntimeError(get_local_runtime_missing_message())
+
+        try:
+            asyncio.get_running_loop()
+        except RuntimeError:
+            pass
+        else:
+            raise RuntimeError(
+                "LangGraphRunner.run() cannot be called from a running event loop. "
+                "Use 'await LangGraphRunner.arun(...)' instead."
+            )
+
+        coro = self._arun_internal(
+            agent=agent,
+            message=message,
+            verbose=verbose,
+            runtime_config=runtime_config,
+            chat_history=chat_history,
+            swallow_aip_logs=swallow_aip_logs,
+            **kwargs,
+        )
+
+        return asyncio.run(coro)
+
+    async def arun(
+        self,
+        agent: Agent,
+        message: str,
+        verbose: bool = False,
+        runtime_config: dict[str, Any] | None = None,
+        chat_history: list[dict[str, str]] | None = None,
+        *,
+        swallow_aip_logs: bool = True,
+        **kwargs: Any,
+    ) -> str:
+        """Execute agent asynchronously and return final response text.
+
+        Args:
+            agent: The glaip_sdk Agent to execute.
+            message: The user message to send to the agent.
+            verbose: If True, emit debug trace output during execution.
+                Defaults to False.
+            runtime_config: Optional runtime configuration for tools, MCPs, etc.
+                Defaults to None. (Implemented in PR-04+)
+            chat_history: Optional list of prior conversation messages.
+                Each message is a dict with "role" and "content" keys.
+                Defaults to None.
+            swallow_aip_logs: When True (default), silence noisy AIPAgents logs.
+            **kwargs: Additional keyword arguments passed to the backend.
+
+        Returns:
+            The final response text from the agent.
+
+        Raises:
+            RuntimeError: If no final response is received from the agent.
+        """
+        return await self._arun_internal(
+            agent=agent,
+            message=message,
+            verbose=verbose,
+            runtime_config=runtime_config,
+            chat_history=chat_history,
+            swallow_aip_logs=swallow_aip_logs,
+            **kwargs,
+        )
+
+    async def _arun_internal(
+        self,
+        agent: Agent,
+        message: str,
+        verbose: bool = False,
+        runtime_config: dict[str, Any] | None = None,
+        chat_history: list[dict[str, str]] | None = None,
+        *,
+        swallow_aip_logs: bool = True,
+        **kwargs: Any,
+    ) -> str:
+        """Internal async implementation of agent execution.
+
+        Args:
+            agent: The glaip_sdk Agent to execute.
+            message: The user message to send to the agent.
+            verbose: If True, emit debug trace output during execution.
+            runtime_config: Optional runtime configuration for tools, MCPs, etc.
+            chat_history: Optional list of prior conversation messages.
+            swallow_aip_logs: When True (default), silence noisy AIPAgents logs.
+            **kwargs: Additional keyword arguments passed to the backend.
+
+        Returns:
+            The final response text from the agent.
+        """
+        # Optionally swallow noisy AIPAgents logs
+        if swallow_aip_logs:
+            _swallow_aip_logs()
+
+        # POC/MVP: Create pause/resume callback for interactive HITL input
+        pause_resume_callback = PauseResumeCallback()
+
+        # Build the local LangGraphReactAgent from the glaip_sdk Agent
+        local_agent = self.build_langgraph_agent(
+            agent, runtime_config=runtime_config, pause_resume_callback=pause_resume_callback
+        )
+
+        # Convert chat history to LangChain messages for the agent
+        langchain_messages = _convert_chat_history_to_messages(chat_history)
+        if langchain_messages:
+            kwargs["messages"] = langchain_messages
+            logger.debug(
+                "Passing %d chat history messages to agent '%s'",
+                len(langchain_messages),
+                agent.name,
+            )
+
+        # Use shared render manager for unified processing
+        render_manager = AgentRunRenderingManager(logger)
+        renderer = render_manager.create_renderer(kwargs.get("renderer"), verbose=verbose)
+
+        # POC/MVP: Set renderer on callback so LocalPromptHandler can pause/resume Live
+        pause_resume_callback.set_renderer(renderer)
+
+        meta = render_manager.build_initial_metadata(agent.name, message, kwargs)
+        render_manager.start_renderer(renderer, meta)
+
+        try:
+            # Use shared async stream processor for unified event handling
+            (
+                final_text,
+                stats_usage,
+                started_monotonic,
+                finished_monotonic,
+            ) = await render_manager.async_process_stream_events(
+                local_agent.arun_sse_stream(message, **kwargs),
+                renderer,
+                meta,
+                skip_final_render=True,
+            )
+        except KeyboardInterrupt:
+            try:
+                renderer.close()
+            finally:
+                raise
+        except Exception:
+            try:
+                renderer.close()
+            finally:
+                raise
+
+        # Use shared finalizer to avoid code duplication
+        from glaip_sdk.client.run_rendering import finalize_render_manager  # noqa: PLC0415
+
+        return finalize_render_manager(
+            render_manager, renderer, final_text, stats_usage, started_monotonic, finished_monotonic
+        )
+
+    def build_langgraph_agent(
+        self,
+        agent: Agent,
+        runtime_config: dict[str, Any] | None = None,
+        shared_tool_output_manager: Any | None = None,
+        *,
+        pause_resume_callback: Any | None = None,
+    ) -> Any:
+        """Build a LangGraphReactAgent from a glaip_sdk Agent definition.
+
+        Args:
+            agent: The glaip_sdk Agent to convert.
+            runtime_config: Optional runtime configuration with tool_configs,
+                mcp_configs, agent_config, and agent-specific overrides.
+            shared_tool_output_manager: Optional ToolOutputManager to reuse across
+                agents with tool_output_sharing enabled.
+            pause_resume_callback: Optional callback used to pause/resume the renderer
+                during interactive HITL prompts.
+
+        Returns:
+            A configured LangGraphReactAgent instance.
+
+        Raises:
+            ImportError: If aip-agents is not installed.
+            ValueError: If agent has unsupported tools, MCPs, or sub-agents for local mode.
+        """
+        from aip_agents.agent import LangGraphReactAgent  # noqa: PLC0415
+
+        from glaip_sdk.runner.tool_adapter import LangChainToolAdapter  # noqa: PLC0415
+
+        # Adapt tools for local execution
+        # NOTE: CLI parity waiver - local tool execution is SDK-only for MVP.
+        # See specs/f/local-agent-runtime/plan.md: "CLI parity is explicitly deferred
+        # and will require SDK Technical Lead sign-off per constitution principle IV."
+        langchain_tools: list[Any] = []
+        if agent.tools:
+            adapter = LangChainToolAdapter()
+            langchain_tools = adapter.adapt_tools(agent.tools)
+
+        # Normalize runtime config: merge global and agent-specific configs
+        normalized_config = self._normalize_runtime_config(runtime_config, agent)
+
+        # Merge tool_configs: agent definition < runtime config
+        tool_configs = self._merge_tool_configs(agent, normalized_config)
+
+        # Merge mcp_configs: agent definition < runtime config
+        mcp_configs = self._merge_mcp_configs(agent, normalized_config)
+
+        # Merge agent_config: agent definition < runtime config
+        merged_agent_config = self._merge_agent_config(agent, normalized_config)
+        agent_config_params, agent_config_kwargs = self._apply_agent_config(merged_agent_config)
+
+        tool_output_manager = self._resolve_tool_output_manager(
+            agent,
+            merged_agent_config,
+            shared_tool_output_manager,
+        )
+
+        # Build sub-agents recursively, sharing tool output manager when enabled.
+        sub_agent_instances = self._build_sub_agents(
+            agent.agents,
+            runtime_config,
+            shared_tool_output_manager=tool_output_manager,
+        )
+
+        # Build the LangGraphReactAgent with tools, sub-agents, and configs
+        local_agent = LangGraphReactAgent(
+            name=agent.name,
+            instruction=agent.instruction,
+            description=agent.description,
+            model=agent.model or self.default_model,
+            tools=langchain_tools,
+            agents=sub_agent_instances if sub_agent_instances else None,
+            tool_configs=tool_configs if tool_configs else None,
+            tool_output_manager=tool_output_manager,
+            **agent_config_params,
+            **agent_config_kwargs,
+        )
+
+        # Add MCP servers if configured
+        self._add_mcp_servers(local_agent, agent, mcp_configs)
+
+        # Inject local HITL manager only if hitl_enabled is True (master switch).
+        # This matches remote behavior: hitl_enabled gates the HITL plumbing.
+        # Tool-level HITL configs are only enforced when hitl_enabled=True.
+        self._inject_hitl_manager(local_agent, merged_agent_config, agent.name, pause_resume_callback)
+
+        logger.debug(
+            "Built local LangGraphReactAgent for agent '%s' with %d tools, %d sub-agents, and %d MCPs",
+            agent.name,
+            len(langchain_tools),
+            len(sub_agent_instances),
+            len(agent.mcps) if agent.mcps else 0,
+        )
+        return local_agent
+
+    def _resolve_tool_output_manager(
+        self,
+        agent: Agent,
+        merged_agent_config: dict[str, Any],
+        shared_tool_output_manager: Any | None,
+    ) -> Any | None:
+        """Resolve tool output manager for local agent execution."""
+        tool_output_sharing_enabled = merged_agent_config.get("tool_output_sharing", False)
+        if not tool_output_sharing_enabled:
+            return None
+        if shared_tool_output_manager is not None:
+            return shared_tool_output_manager
+        return build_tool_output_manager(agent.name, merged_agent_config)
+
+    def _inject_hitl_manager(
+        self,
+        local_agent: Any,
+        merged_agent_config: dict[str, Any],
+        agent_name: str,
+        pause_resume_callback: Any | None,
+    ) -> None:
+        """Inject HITL manager when enabled, mirroring remote gating behavior."""
+        hitl_enabled = merged_agent_config.get("hitl_enabled", False)
+        if hitl_enabled:
+            try:
+                from aip_agents.agent.hitl.manager import ApprovalManager  # noqa: PLC0415
+                from glaip_sdk.hitl import LocalPromptHandler  # noqa: PLC0415
+
+                local_agent.hitl_manager = ApprovalManager(
+                    prompt_handler=LocalPromptHandler(pause_resume_callback=pause_resume_callback)
+                )
+                # Store callback reference for setting renderer later
+                if pause_resume_callback:
+                    local_agent._pause_resume_callback = pause_resume_callback
+                logger.debug("HITL manager injected for agent '%s' (hitl_enabled=True)", agent_name)
+            except ImportError as e:
+                # Missing dependencies - fail fast
+                raise ImportError("Local HITL requires aip_agents. Install with: pip install 'glaip-sdk[local]'") from e
+            except Exception as e:
+                # Other errors during HITL setup - fail fast
+                raise RuntimeError(f"Failed to initialize HITL manager for agent '{agent_name}'") from e
+        else:
+            logger.debug("HITL manager not injected for agent '%s' (hitl_enabled=False)", agent_name)
+
+    def _build_sub_agents(
+        self,
+        sub_agents: list[Any] | None,
+        runtime_config: dict[str, Any] | None,
+        shared_tool_output_manager: Any | None = None,
+    ) -> list[Any]:
+        """Build sub-agent instances recursively.
+
+        Args:
+            sub_agents: List of sub-agent definitions.
+            runtime_config: Runtime config to pass to sub-agents.
+            shared_tool_output_manager: Optional ToolOutputManager to reuse across
+                agents with tool_output_sharing enabled.
+
+        Returns:
+            List of built sub-agent instances.
+
+        Raises:
+            ValueError: If sub-agent is platform-only.
+        """
+        if not sub_agents:
+            return []
+
+        sub_agent_instances = []
+        for sub_agent in sub_agents:
+            self._validate_sub_agent_for_local_mode(sub_agent)
+            sub_agent_instances.append(
+                self.build_langgraph_agent(
+                    sub_agent,
+                    runtime_config,
+                    shared_tool_output_manager=shared_tool_output_manager,
+                )
+            )
+        return sub_agent_instances
+
+    def _add_mcp_servers(
+        self,
+        local_agent: Any,
+        agent: Agent,
+        merged_mcp_configs: dict[str, Any],
+    ) -> None:
+        """Add MCP servers to a built agent.
+
+        Args:
+            local_agent: The LangGraphReactAgent to add MCPs to.
+            agent: The glaip_sdk Agent with MCP definitions.
+            merged_mcp_configs: Merged mcp_configs (agent definition + runtime).
+        """
+        if not agent.mcps:
+            return
+
+        from glaip_sdk.runner.mcp_adapter import LangChainMCPAdapter  # noqa: PLC0415
+
+        mcp_adapter = LangChainMCPAdapter()
+        base_mcp_configs = mcp_adapter.adapt_mcps(agent.mcps)
+
+        # Apply merged mcp_configs overrides (agent definition + runtime)
+        if merged_mcp_configs:
+            base_mcp_configs = self._apply_runtime_mcp_configs(base_mcp_configs, merged_mcp_configs)
+
+        if base_mcp_configs:
+            local_agent.add_mcp_server(base_mcp_configs)
+            logger.debug(
+                "Registered %d MCP server(s) for agent '%s'",
+                len(base_mcp_configs),
+                agent.name,
+            )
+
+    def _normalize_runtime_config(
+        self,
+        runtime_config: dict[str, Any] | None,
+        agent: Agent,
+    ) -> dict[str, Any]:
+        """Normalize runtime_config for local execution.
+
+        Merges global and agent-specific configs with proper priority.
+        Keys are resolved from instances/classes to string names.
+
+        Args:
+            runtime_config: Raw runtime config from user.
+            agent: The agent being built (for resolving agent-specific overrides).
+
+        Returns:
+            Normalized config with string keys and merged priorities.
+        """
+        from glaip_sdk.utils.runtime_config import (  # noqa: PLC0415
+            merge_configs,
+            normalize_local_config_keys,
+        )
+
+        if not runtime_config:
+            return {}
+
+        # 1. Extract global configs and normalize keys
+        global_tool_configs = normalize_local_config_keys(runtime_config.get("tool_configs", {}))
+        global_mcp_configs = normalize_local_config_keys(runtime_config.get("mcp_configs", {}))
+        global_agent_config = runtime_config.get("agent_config", {})
+
+        # 2. Extract agent-specific overrides (highest priority)
+        agent_specific = self._get_agent_specific_config(runtime_config, agent)
+        agent_tool_configs = normalize_local_config_keys(agent_specific.get("tool_configs", {}))
+        agent_mcp_configs = normalize_local_config_keys(agent_specific.get("mcp_configs", {}))
+        agent_config_override = agent_specific.get("agent_config", {})
+
+        # 3. Merge with priority: global < agent-specific
+        merged_result = {
+            "tool_configs": merge_configs(global_tool_configs, agent_tool_configs),
+            "mcp_configs": merge_configs(global_mcp_configs, agent_mcp_configs),
+            "agent_config": merge_configs(global_agent_config, agent_config_override),
+        }
+        return merged_result
+
+    def _get_agent_specific_config(
+        self,
+        runtime_config: dict[str, Any],
+        agent: Agent,
+    ) -> dict[str, Any]:
+        """Extract agent-specific config from runtime_config.
+
+        Args:
+            runtime_config: Runtime config that may contain agent-specific overrides.
+            agent: The agent to find config for.
+
+        Returns:
+            Agent-specific config dict, or empty dict if not found.
+        """
+        from glaip_sdk.utils.resource_refs import is_uuid  # noqa: PLC0415
+        from glaip_sdk.utils.runtime_config import get_name_from_key  # noqa: PLC0415
+
+        # Reserved keys at the top level
+        reserved_keys = {"tool_configs", "mcp_configs", "agent_config"}
+
+        # Try finding agent by instance, class, or name
+        for key, value in runtime_config.items():
+            if key in reserved_keys:
+                continue  # Skip global configs
+
+            if isinstance(key, str) and is_uuid(key):
+                logger.warning(
+                    "UUID agent override key '%s' is not supported in local mode; skipping. "
+                    "Use agent name string or Agent instance as the key instead.",
+                    key,
+                )
+                continue
+
+            # Check if this key matches the agent
+            try:
+                key_name = get_name_from_key(key)
+            except ValueError:
+                continue  # Skip invalid keys
+
+            if key_name and key_name == agent.name:
+                return value if isinstance(value, dict) else {}
+
+        return {}
+
+    def _merge_tool_configs(
+        self,
+        agent: Agent,
+        normalized_config: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Merge agent.tool_configs with runtime tool_configs.
+
+        Priority (lowest to highest):
+        1. Agent definition (agent.tool_configs)
+        2. Runtime config (normalized_config["tool_configs"])
+
+        Args:
+            agent: The agent with optional tool_configs property.
+            normalized_config: Normalized runtime config.
+
+        Returns:
+            Merged tool_configs dict.
+        """
+        from glaip_sdk.utils.runtime_config import (  # noqa: PLC0415
+            merge_configs,
+            normalize_local_config_keys,
+        )
+
+        # Get agent's tool_configs if defined
+        agent_tool_configs = {}
+        if hasattr(agent, "tool_configs") and agent.tool_configs:
+            agent_tool_configs = normalize_local_config_keys(agent.tool_configs)
+
+        # Get runtime tool_configs
+        runtime_tool_configs = normalized_config.get("tool_configs", {})
+
+        # Merge: agent definition < runtime config
+        return merge_configs(agent_tool_configs, runtime_tool_configs)
+
+    def _merge_mcp_configs(
+        self,
+        agent: Agent,
+        normalized_config: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Merge agent.mcp_configs with runtime mcp_configs.
+
+        Priority (lowest to highest):
+        1. Agent definition (agent.mcp_configs)
+        2. Runtime config (normalized_config["mcp_configs"])
+
+        Args:
+            agent: The agent with optional mcp_configs property.
+            normalized_config: Normalized runtime config.
+
+        Returns:
+            Merged mcp_configs dict.
+        """
+        from glaip_sdk.utils.runtime_config import (  # noqa: PLC0415
+            merge_configs,
+            normalize_local_config_keys,
+        )
+
+        # Get agent's mcp_configs if defined
+        agent_mcp_configs = {}
+        if hasattr(agent, "mcp_configs") and agent.mcp_configs:
+            agent_mcp_configs = normalize_local_config_keys(agent.mcp_configs)
+
+        # Get runtime mcp_configs
+        runtime_mcp_configs = normalized_config.get("mcp_configs", {})
+
+        # Merge: agent definition < runtime config
+        return merge_configs(agent_mcp_configs, runtime_mcp_configs)
+
+    def _merge_agent_config(
+        self,
+        agent: Agent,
+        normalized_config: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Merge agent.agent_config with runtime agent_config.
+
+        Priority (lowest to highest):
+        1. Agent definition (agent.agent_config)
+        2. Runtime config (normalized_config["agent_config"])
+
+        Args:
+            agent: The agent with optional agent_config property.
+            normalized_config: Normalized runtime config.
+
+        Returns:
+            Merged agent_config dict.
+        """
+        from glaip_sdk.utils.runtime_config import merge_configs  # noqa: PLC0415
+
+        # Get agent's agent_config if defined
+        agent_agent_config = {}
+        if hasattr(agent, "agent_config") and agent.agent_config:
+            agent_agent_config = agent.agent_config
+
+        # Get runtime agent_config
+        runtime_agent_config = normalized_config.get("agent_config", {})
+
+        # Merge: agent definition < runtime config
+        return merge_configs(agent_agent_config, runtime_agent_config)
+
+    def _apply_agent_config(
+        self,
+        agent_config: dict[str, Any],
+    ) -> tuple[dict[str, Any], dict[str, Any]]:
+        """Extract and separate agent_config into direct params and kwargs.
+
+        Separates agent_config into parameters that go directly to LangGraphReactAgent
+        constructor vs those that go through **kwargs.
+
+        Args:
+            agent_config: Runtime agent configuration dict.
+
+        Returns:
+            Tuple of (direct_params, kwargs_params):
+            - direct_params: Parameters passed directly to LangGraphReactAgent.__init__()
+            - kwargs_params: Parameters passed via **kwargs to BaseAgent
+        """
+        direct_params = {}
+        kwargs_params = {}
+
+        # Direct constructor parameters
+        if "planning" in agent_config:
+            direct_params["planning"] = agent_config["planning"]
+
+        if "enable_a2a_token_streaming" in agent_config:
+            direct_params["enable_a2a_token_streaming"] = agent_config["enable_a2a_token_streaming"]
+
+        # Kwargs parameters (passed through **kwargs to BaseAgent)
+        if "enable_pii" in agent_config:
+            kwargs_params["enable_pii"] = agent_config["enable_pii"]
+
+        if "memory" in agent_config:
+            # Map "memory" to "memory_backend" for aip-agents compatibility
+            kwargs_params["memory_backend"] = agent_config["memory"]
+
+        # Additional memory-related settings
+        memory_settings = ["agent_id", "memory_namespace", "save_interaction_to_memory"]
+        for key in memory_settings:
+            if key in agent_config:
+                kwargs_params[key] = agent_config[key]
+
+        return direct_params, kwargs_params
+
+    def _apply_runtime_mcp_configs(
+        self,
+        base_configs: dict[str, Any],
+        runtime_overrides: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Apply runtime mcp_configs overrides to base MCP configurations.
+
+        Merges runtime overrides into the base configs, handling authentication
+        conversion to headers using MCPConfigBuilder.
+
+        Args:
+            base_configs: Base MCP configs from adapter (server_name -> config).
+            runtime_overrides: Runtime mcp_configs overrides (server_name -> config).
+
+        Returns:
+            Merged MCP configs with authentication converted to headers.
+        """
+        return {
+            server_name: self._merge_single_mcp_config(server_name, base_config, runtime_overrides.get(server_name))
+            for server_name, base_config in base_configs.items()
+        }
+
+    def _merge_single_mcp_config(
+        self,
+        server_name: str,
+        base_config: dict[str, Any],
+        override: dict[str, Any] | None,
+    ) -> dict[str, Any]:
+        """Merge a single MCP config with runtime override.
+
+        Args:
+            server_name: Name of the MCP server.
+            base_config: Base config from adapter.
+            override: Optional runtime override config.
+
+        Returns:
+            Merged config dict.
+        """
+        merged = base_config.copy()
+
+        if not override:
+            return merged
+
+        from glaip_sdk.runner.mcp_adapter.mcp_config_builder import (  # noqa: PLC0415
+            MCPConfigBuilder,
+        )
+
+        # Handle authentication override
+        if "authentication" in override:
+            headers = MCPConfigBuilder.build_headers_from_auth(override["authentication"])
+            if headers:
+                merged["headers"] = headers
+                logger.debug("Applied runtime authentication headers for MCP '%s'", server_name)
+
+        # Merge other config keys (excluding authentication since we converted it)
+        for key, value in override.items():
+            if key != "authentication":
+                merged[key] = value
+
+        return merged
+
+    def _validate_sub_agent_for_local_mode(self, sub_agent: Any) -> None:
+        """Validate that a sub-agent reference is supported for local execution.
+
+        Args:
+            sub_agent: The sub-agent reference to validate.
+
+        Raises:
+            ValueError: If the sub-agent is not supported in local mode.
+        """
+        # String references are allowed by SDK API but not for local mode
+        if isinstance(sub_agent, str):
+            raise ValueError(
+                f"Sub-agent '{sub_agent}' is a string reference and cannot be used in local mode. "
+                "String sub-agent references are only supported for server execution. "
+                "For local mode, define the sub-agent with Agent(name=..., instruction=...)."
+            )
+
+        # Validate sub-agent is not a class
+        if inspect.isclass(sub_agent):
+            raise ValueError(
+                f"Sub-agent '{sub_agent.__name__}' is a class, not an instance. "
+                "Local mode requires Agent INSTANCES. "
+                "Did you forget to instantiate it? e.g., Agent(...), not Agent"
+            )
+
+        # Validate sub-agent is an Agent-like object (has required attributes)
+        if not hasattr(sub_agent, "name") or not hasattr(sub_agent, "instruction"):
+            raise ValueError(
+                f"Sub-agent {type(sub_agent).__name__} is not supported in local mode. "
+                "Local mode requires Agent instances with 'name' and 'instruction' attributes. "
+                "Define the sub-agent with Agent(name=..., instruction=...)."
+            )
+
+        # Validate sub-agent is not platform-only (from_id, from_native)
+        if getattr(sub_agent, "_lookup_only", False):
+            agent_name = getattr(sub_agent, "name", "<unknown>")
+            raise ValueError(
+                f"Sub-agent '{agent_name}' is not supported in local mode. "
+                "Platform agents (from_id, from_native) cannot be used as "
+                "sub-agents in local execution. "
+                "Define the sub-agent locally with Agent(name=..., instruction=...) instead."
+            )
+
+    def _log_event(self, event: dict[str, Any]) -> None:
+        """Log an A2AEvent for verbose debug output.
+
+        Args:
+            event: The A2AEvent dictionary to log.
+        """
+        event_type = event.get("event_type", "unknown")
+        content = event.get("content", "")
+        is_final = event.get("is_final", False)
+
+        # Truncate long content for readability
+        content_str = str(content) if content else ""
+        content_preview = content_str[:100] + "..." if len(content_str) > 100 else content_str
+
+        final_marker = "(final)" if is_final else ""
+        logger.info("[%s] %s %s", event_type, final_marker, content_preview)