fix: optional mcp tools

[NicoleMGomes]

Disclaimer: Do not include SAP-internal or customer-specific information in this PR (e.g. internal system URLs, customer names, tenant IDs, or confidential configurations). This is a public repository.

Description

Fixes a bug in mcp_tool_to_langchain (agentgateway/converters.py) where all MCP tool parameters were unconditionally marked as required in the generated Pydantic args_schema, regardless of what the tool's input_schema declared as optional.

Root cause: The field mapping {k: (str, ...) for k in properties} used the Pydantic sentinel ... (required) for every field. Fields not listed in the JSON Schema "required" array were never treated as optional.

Fix: The converter now reads the "required" array from input_schema and maps fields accordingly — fields in "required" remain (str, ...), while all other fields become (str | None, None) (optional, defaulting to None).

This means LangChain agents and any Pydantic-aware consumer using mcp_tool_to_langchain will no longer raise validation errors when optional MCP tool parameters are omitted.

Related Issue

Closes #

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update
• Code refactoring
• Dependency update

How to Test

1. Define an MCPTool whose input_schema has both required and optional properties:

   tool = MCPTool(
       name="get_supplier_bid",
       server_name="ariba",
       description="Gets supplier bids",
       input_schema={
           "type": "object",
           "required": ["eventid"],
           "properties": {
               "eventid": {"description": "Event ID"},
               "datafetchmode": {"description": "Detail level"},
           },
       },
       url="https://example.com/mcp",
   )

2. Convert it: langchain_tool = mcp_tool_to_langchain(tool, call_tool, get_user_token)
3. Inspect langchain_tool.args_schema.model_fields:
   • eventid must be required (is_required() == True)
   • datafetchmode must be optional (is_required() == False)
4. Run the unit tests: pytest tests/agentgateway/unit/test_converters.py -v

Checklist

• I have read the Contributing Guidelines
• I have verified that my changes solve the issue
• I have added/updated automated tests to cover my changes
• All tests pass locally
• I have verified that my code follows the Code Guidelines
• I have updated documentation (if applicable)
• I have added type hints for all public APIs
• My code does not contain sensitive information (credentials, tokens, etc.)
• I have followed Conventional Commits for commit messages

Additional Notes

The fix is confined to the mcp_tool_to_langchain reference converter. The core AgentGatewayClient.call_mcp_tool() is unaffected — it still passes **kwargs through to the MCP server without schema validation.

[thiagob]

lgm