[adityasingh2400]

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

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

• Closes: Failed MCP tool results are sent to the Interactions API with is_error=False #7824

Problem:
_convert_part_to_interaction_content set is_error on a function_result step only when the function response had an error key. A failed MCP CallToolResult reports the failure with isError (is_error in MCP SDK 2.x dumps) and has no error key, so on the Interactions API Gemini was told that a failed MCP tool call succeeded.

Solution:
Also treat a truthy isError or is_error in the response dict as an error. This matches what AnthropicLlm does since #7665. Responses with an error key behave as before, and the result payload is still passed through unchanged.

Testing Plan

Unit Tests:

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

Added to tests/unittests/models/test_interactions_utils.py, each parametrized over isError and is_error:

• test_function_response_failed_mcp_result_is_error
• test_function_response_successful_mcp_result_is_not_error

# On main (before the fix)
test_function_response_failed_mcp_result_is_error[isError]   assert False is True
test_function_response_failed_mcp_result_is_error[is_error]  assert False is True

# With the fix
$ pytest tests/unittests/models/test_interactions_utils.py tests/unittests/models/test_google_llm.py
335 passed

pre-commit run passes on the changed files. mypy reports one error in interactions_utils.py (FunctionResultDelta has no attribute call_id), which is already on main and outside this change.

Manual End-to-End (E2E) Tests:

The test server is a real streamable-HTTP MCP server whose read_file tool raises, so it returns an isError: true result. A Runner drives an LlmAgent(model=Gemini("gemini-2.5-flash", use_interactions_api=True), tools=[McpToolset(StreamableHTTPConnectionParams(...))]). Only interactions_utils._create_interactions (the network call) is stubbed: it records create_kwargs["input"], asks for read_file on the first request, and answers on the second. The output below is the function_result step from the second request:

=== BEFORE (main) ===
function_result sent to Gemini: name=read_file is_error=False result='Error executing tool read_file'
=== AFTER (this PR) ===
function_result sent to Gemini: name=read_file is_error=True result='Error executing tool read_file'

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

None.

[adk-bot]

Thank you @adityasingh2400 for your contribution! 🎉

Your changes have been successfully imported and merged via Copybara in commit 72205cd.

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