FoundationAgents/OpenManus · error · ServerError

Error invoking agent: {e}

Error message

Error invoking agent: {e}

What it means

Raised by the A2A agent executor when constructing or invoking the agent fails: any exception from agent_factory() or agent.invoke() (including the NotImplementedError from stream-based misuse, network errors to backing LLMs, missing API keys, or agent timeouts) is wrapped in ServerError(ValueError('Error invoking agent: ...')). The original exception is chained via 'from e'.

Source

Thrown at protocol/a2a/app/agent_executor.py:45

        self.agent_factory = agent_factory

    async def execute(
        self,
        context: RequestContext,
        event_queue: EventQueue,
    ) -> None:
        error = self._validate_request(context)
        if error:
            raise ServerError(error=InvalidParamsError())

        query = context.get_user_input()
        try:
            self.agent = await self.agent_factory()
            result = await self.agent.invoke(query, context.context_id)
            print(f"Final Result ===> {result}")
        except Exception as e:
            print("Error invoking agent: %s", e)
            raise ServerError(error=ValueError(f"Error invoking agent: {e}")) from e
        parts = [
            Part(
                root=TextPart(
                    text=(
                        result["content"]
                        if result["content"]
                        else "failed to generate response"
                    )
                ),
            )
        ]
        event_queue.enqueue_event(
            completed_task(
                context.task_id,
                context.context_id,
                [new_artifact(parts, f"task_{context.task_id}")],
                [context.message],
            )

View on GitHub (pinned to 52a13f2a57)

Solutions

  1. Check the chained exception (raise ... from e preserves it) — print e.__cause__ or enable traceback logging to see the root cause.
  2. Verify the agent's runtime prerequisites: API keys set, network egress allowed, factory callable returns a healthy agent.
  3. Test the agent directly (await agent.invoke(query, sid)) outside the executor to isolate executor vs agent failure.
  4. Fix the logging bug: use print(f'Error invoking agent: {e}') or logging.exception(...) so the actual error appears in server logs.

Example fix

# before
except Exception as e:
    print("Error invoking agent: %s", e)  # literal '%s' in output, cause hidden
    raise ServerError(error=ValueError(f"Error invoking agent: {e}")) from e
# after
except Exception as e:
    logging.exception("Error invoking agent")  # full traceback including __cause__
    raise ServerError(error=ValueError(f"Error invoking agent: {e}")) from e
Defensive patterns

Strategy: try-catch

Validate before calling

# verify agent prerequisites before serving requests
assert os.environ.get('LLM_API_KEY'), 'missing LLM_API_KEY'
agent = await agent_factory()  # fail fast at startup, not per request

Try / catch

try:
    result = await agent.invoke(query, context.context_id)
except ServerError as e:
    root = e.error.message if e.error else str(e)
    if root.startswith('Error invoking agent:'):
        cause = e.__cause__ or e  # original exception is chained
        log.error('agent failure: %r', cause)
        # surface cause details, do not retry blindly

Prevention

When it happens

Trigger: agent_factory() raising (bad config, missing env vars like API keys, dependency import errors), or agent.invoke() raising mid-run (LLM provider error, tool failure inside Manus, invalid session/context). The printed line uses %s inside print(), which is a formatting bug — the raw '%s' placeholder and exception print literally, so logs look odd.

Common situations: Missing or expired LLM API credentials in the server environment; sandboxed deployments blocking outbound LLM calls; misconfigured agent factory; the underlying agent hitting its own internal error. The poor log line ('Error invoking agent: %s', e) often hides the real cause, forcing developers to debug blind.

Related errors

AI-assisted analysis of FoundationAgents/OpenManus@52a13f2a57 (2026-08-15). Data as JSON: /api/errors/98eb5d104dea7541. Report an issue: GitHub.