Worked on the langchain-ai/langchain repository to enhance the robustness of tool argument handling within Langchain integrations. Developed a feature that enables automatic injection of Langchain-specific arguments, such as InjectedToolCallId and ToolRuntime, even when custom args_schema definitions omit them. This approach, implemented using Python and leveraging Pydantic for schema validation, ensures that argument processing remains consistent and reliable across diverse tool configurations. By improving the flexibility of argument injection and validation, the work reduced integration edge-cases and runtime errors, laying the groundwork for broader schema support and enabling more complex workflows within the Langchain Core ecosystem.