fix(python): run basic ADK sample through kagent-adk

[tryambak2019]

Summary

Fix the ADK basic sample container command and validate its supported runtime-loading path.

Implemented

• Run the sample through kagent-adk run basic from its package directory.
• Add CI coverage that loads the real root_agent through the ADK CLI with Uvicorn mocked.
• Preserve Google ADK session semantics; the test uses the CLI's explicit local in-memory mode.

Validation

• Focused CLI regression test passed.
• make -C python test passed.
• Ruff lint, format, and uv lock --check passed.
• Docker build is blocked before execution because the default external kagent-adk:latest base image is not published in GHCR.

#2689 Series

• Documentation corrections: docs(python): align integration READMEs with API v2 #2821
• Currency entry-point validation: fix(python): make currency CLI entry point importable #2829
• Basic OpenAI entry-point validation: fix(python): make basic OpenAI agent entry point importable #2830
• LangGraph sample entry-point validation: fix(python): make LangGraph sample entry points importable #2831
• ADK basic runtime-command validation: this PR
• Remaining AgentSTS audit, CrewAI validation, and residual cleanup: Revisit Python packages and samples after API v2 cutover #2689

Related to #2689. This PR does not close it.

[supreme-gg-gg]

@tryambak2019 Your "series" of PR is very difficult to review, can you put all Python changes into one single PR since they're related to the same issue. Especially that I found a few possible issues with pytest and health checks that will become apparent once they're combined. This also reduces overhead for merge conflicts and CI runs. I'll take another look once that's done

For the docs PR #2821, if you make its doc changes independent of your new python changes, then I think it can stay separate