[BlueRaddish]

Please ensure you have read the contribution guide before creating a pull request.

Link to Issue or Description of Change

Problem: InMemoryRunner.run_debug(verbose=False) promises output showing only final agent responses, but the shared event printer includes Part(thought=True) text. A response containing a thought followed by an answer is printed as test_agent > thinking answer; a thought-only event is also printed.

Solution: Skip thought parts in nonverbose output. Preserve the existing verbose output for debugging and retain all original parts in the events returned by run_debug. No model configuration, session persistence, or event serialization changes.

Reproduction: A deterministic BaseAgent yields an Event with a thought part followed by an answer part. Running it with the public InMemoryRunner.run_debug entry point prints the thought on the base revision; the same invocation with this change prints only test_agent > answer. No inference service or external network is involved.

Testing Plan

Unit Tests:

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

python -m pytest tests/unittests/utils/test_debug_output.py tests/unittests/test_runners.py -k 'print_event or run_debug' -q

Result: 20 passed, 182 deselected. Before the fix, the printer tests had 2 failed, 15 passed, and the public runner tests had 1 failed, 2 passed. Coverage includes a mixed thought/answer event, a thought-only event, verbose output, and preservation of the thought in returned events. The full ADK suite was not run.

Manual End-to-End (E2E) Tests: The two parameterized test_run_debug_hides_thoughts_unless_verbose cases execute the real public runner, a deterministic custom BaseAgent, and real in-memory session services. They exercise both console modes without mocking the printer or running a model. The nonverbose case failed on the base revision and passes after the fix.

All repository pre-commit hooks passed on the exact diff in a native Linux checkout of the same base revision. Windows tests use Python 3.13; no paid compute or model calls were used.

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.
• Relevant new and existing unit tests pass locally with my changes.
• I have tested the change through the public runner end-to-end.
• There are no dependent downstream changes.

Additional context

Prepared with assistance from OpenAI Codex. This is a console verbosity fix, not a claim that debug output is a security boundary. Verbose mode intentionally continues to display the existing thought text.

[adk-bot]

Thank you @BlueRaddish for your contribution! 🎉

Your changes have been successfully imported and merged via Copybara in commit 53cb763.

Closing this PR as the changes are now in the main branch.