Feature/google bridge 2

[anthonyduong9]

This PR contains:

• New features
• Changes to dev-tools e.g. CI config / github tooling
• Docs
• Bug fixes
• Code refactor

What is the current behavior? (You can also link to an open issue here)

Users cannot run Gemini CLI inside (or outside) of a sandbox, using Inspect.

What is the new behavior?

Users cannot run Gemini CLI inside (or outside) of a sandbox, using Inspect.

Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?)

No.

Other information:

[anthonyduong9]

@jjallaire I ran

• pytest src/inspect_sandbox_tools/tests/agent_bridge/test_model_proxy.py
• pytest tests/agent/test_agent_sandbox_bridge.py::test_agent_sandbox_bridge_google --runslow --runapi

and all tests passed.

I also ran pytest tests/agent/test_agent_bridge.py --runapi and the output was

PASSED tests/agent/test_agent_bridge.py::test_bridge_filter_output
PASSED tests/agent/test_agent_bridge.py::test_bridge_filter_config
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_completions
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_completions_tools
PASSED tests/agent/test_agent_bridge.py::test_bridged_web_search_tool_openai
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_anthropic
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_anthropic_tools
PASSED tests/agent/test_agent_bridge.py::test_bridged_web_search_tool_anthropic
PASSED tests/agent/test_agent_bridge.py::test_bridged_web_search_tool_openai_to_anthropic
PASSED tests/agent/test_agent_bridge.py::test_bridged_web_search_tool_anthropic_to_openai
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_google
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_google_tools
PASSED tests/agent/test_agent_bridge.py::test_bridged_web_search_tool_google
PASSED tests/agent/test_agent_bridge.py::test_anthropic_bridged_agent
PASSED tests/agent/test_agent_bridge.py::test_bridged_agent_context
FAILED tests/agent/test_agent_bridge.py::test_bridged_agent_responses - AssertionError: assert 'error' == 'success'
FAILED tests/agent/test_agent_bridge.py::test_bridged_agent_responses_tools - AssertionError: assert 'error' == 'success'

.
Those two tests were already failing on main for me, I think because of my local setup.

My google.genai version is also 1.44.0.