fix(flows): preserve code execution continuation

[saransh-translucent]

Link to Issue or Description of Change

• Closes: Code execution continuation is misclassified as MODEL_RETURNED_NO_CONTENT #7002

Problem:

The non-built-in code-execution response processor intentionally clears llm_response.content after emitting a sandbox result so the LLM loop continues. The genuine-empty-STOP guard currently runs after response processors, mistakes that internal continuation sentinel for a provider-empty response, and emits MODEL_RETURNED_NO_CONTENT before the model can produce its final answer.

Solution:

Classify genuine non-streaming empty STOP responses before response processors run. This keeps the existing provider-empty error while allowing processor-cleared code-execution responses to continue the loop.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

$ .venv/bin/pytest tests/unittests/flows/llm_flows/test_base_llm_flow.py -q
101 passed, 8 warnings in 2.55s

The new regression test uses a deterministic two-response model and a recording non-built-in executor. It verifies that the code executes, a second model call occurs, the final response is emitted, and no MODEL_RETURNED_NO_CONTENT event appears. The existing genuine-empty response regression test remains green.

Manual End-to-End (E2E) Tests:

The same test drives the public in-memory Runner through the complete model → code executor → execution result → second model call → final response flow without external model or customer data.

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.

Additional context

This is a processor-ordering fix. It does not suppress genuine empty model responses and does not change SSE terminal-chunk handling.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.