Fix #61 Continue agent execution when function calls are pending

[meganetaaan]

This fixes #61.

There is an inconsistency between the Python and TypeScript implementations of the final output logic in the agent runner. The Python version correctly checks if tools are running before treating the output as final, while the TypeScript version does not.

Python

elif (
    not output_schema or output_schema.is_plain_text()
) and not processed_response.has_tools_or_approvals_to_run():
    # Only treat as final output if NO tools are running
    return await cls.execute_final_output(...)

TypeScript

if (agent.outputType === 'text') {
  // Missing check for hasToolsOrApprovalsToRun()
  return new SingleStepResult(/* next_step_final_output */);
}

The TypeScript version lacks the crucial hasToolsOrApprovalsToRun() check, causing any text output to be treated as final even when tools are present.

Fix Implemented

Added missing hasToolsOrApprovalsToRun() method to ProcessedResponse and updated final output logic:

if (agent.outputType === 'text' && !processedResponse.hasToolsOrApprovalsToRun()) {
  return new SingleStepResult(/* next_step_final_output */);
}

This aligns the TypeScript implementation with the Python version, allowing streams to continue after tool execution when appropriate.

[DroopyTersen]

I think this might be related to #57

[dkundel-openai]

Hey @meganetaaan thank you for your PR! Looking at this PR and @DroopyTersen PR (thank you for that), I think this aligns closer to the internal Python Agents SDK so I'd like to merge this. Could you please run pnpm changeset to add a patch change to this PR? Also any additional tests would be helpful. Thank you

[changeset-bot]

🦋 Changeset detected

Latest commit: b53cfd9

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 5 packages

Name	Type
@openai/agents-core	Patch
@openai/agents-openai	Patch
@openai/agents-realtime	Patch
@openai/agents	Patch
@openai/agents-extensions	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[meganetaaan]

Thanks for the feedback! I've addressed both requests:

• Added pnpm changeset (patch change for @openai/agents-core)
• Added comprehensive tests for the executeToolsAndSideEffects behavior and
  hasToolsOrApprovalsToRun method

The tests verify that text agents continue execution when tools are pending and return final output
when no tools are pending, which covers the core functionality of this fix.