Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.narrative.io/llms.txt

Use this file to discover all available pages before exploring further.

HTTP status: 400

​
Error message

tool_use_id(s) refer to server-resolved calls (no caller output expected): {ids}

​
When this error occurs

Each tool call the model emits is routed by its alias prefix:
  • An alias listed under mcp_servers is server-resolved — the workflow calls the MCP server inline and persists the result; you never see those in pending_tool_calls.
  • An alias listed under client_tools is client-resolved — the run pauses at requires_action and the caller posts the output as tool_outputs on the next run.
This error means a tool_outputs[].tool_use_id references an assistant turn block whose alias is in mcp_servers, not client_tools. The server already produced (or failed to produce) the result for that call; there’s nothing for the caller to contribute.

​
How to fix

Filter your tool_outputs to only the ids whose name in the prior requires_action response started with a client_tools alias. The pending_tool_calls list from GET /agents/runs/{id} already excludes server-resolved ids — copy from there rather than re-deriving from the message stream.