fix(tools): forward custom name and description from VertexAiSearchTool bypass path to DiscoveryEngineSearchTool

[musi22]

Description

Addresses Problem 3 identified in #7100.

When bypass_multi_tools_limit=True is enabled on VertexAiSearchTool, ADK converts the tool to a client-side DiscoveryEngineSearchTool during canonical tool resolution. Previously, this conversion always named the resulting tool 'discovery_engine_search' -- an internal implementation detail.

This caused two key problems for developers:

1. Instruction coupling: Agents had to reference 'discovery_engine_search' explicitly in prompts rather than domain concepts like 'product_search' or 'knowledge_base'.
2. 2. Runtime crashes: Lightweight or third-party LLMs frequently guess natural tool names (like search or query) instead of discovery_engine_search, leading to ValueError: Tool '...' not found.
      While DiscoveryEngineSearchTool already supported name and description parameters, they were unreachable through the VertexAiSearchTool conversion path.

Changes

• Added optional name and description parameters to VertexAiSearchTool.__init__.
• Stored them internally as _bypass_tool_name and _bypass_tool_description to avoid shadowing BaseTool.name = 'vertex_ai_search' (preserving backward compatibility with the built-in grounding path).
• Forwarded name and description during bypass conversion in llm_agent.py.
• Added 5 unit tests in tests/unittests/agents/test_llm_agent_fields.py.
• Added 4 unit tests in tests/unittests/tools/test_discovery_engine_search_tool.py.

Related Issues

• Partially fixes bypass_multi_tools_limit on VertexAiSearchTool is a leaky abstraction: unexpected dependencies, silent shift from inbuilt RAG to tool calling, and prompt fragility #7100 (Problem 3: Hardcoded tool name)
• Non-overlapping with fix: raise clear ImportError when VertexAiSearchTool bypass needs google-adk[gcp] #7101 (which focuses strictly on missing google-cloud-discoveryengine import errors)

Testing Plan

• Ran unit tests for DiscoveryEngineSearchTool (pytest tests/unittests/tools/test_discovery_engine_search_tool.py) -- 31/31 passed.
  -[x] Ran unit tests for VAIS tool conversion (pytest tests/unittests/agents/test_llm_agent_fields.py -k "vais") -- 10/10 passed.

[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.

[musi22]

@googlebot I signed it!

[musi22]

Just to clarify the scope of this PR: it intentionally addresses only Problem 3 from #7100—the hardcoded name and description used when VertexAiSearchTool is converted to DiscoveryEngineSearchTool.

The change allows developers to provide a domain-specific tool name and description through VertexAiSearchTool, while preserving the existing vertex_ai_search identity and behavior for the normal built-in grounding path.

I kept this PR deliberately focused. It does not change the underlying retrieval architecture, add fuzzy tool-name matching, or address the optional GCP dependency behavior.

I also noticed that #7121 overlaps with this area. I’m happy to adjust, rebase, or narrow this PR further depending on which implementation direction the maintainers prefer. The focused conversion and constructor tests are included and passing.

Please let me know if any additional clarification or test coverage would help with the review.