Python AI Agent Developer - Tool-Calling Reliability in an Ongoing Project

We're working on an ongoing Python-based AI assistant that connects to external services through API tools. The core workflow is already in place, but we've run into an issue that needs a focused fix before we continue development. The agent can trigger the correct tool and receive a response, yet it sometimes ignores that response when generating its answer or asks for information the tool has already returned. We need someone to trace where the result is being lost and correct the handling. For this task, you’ll: - Review the tool-call and response-handling flow. - Fix the identified issue without restructuring the application. - Handle empty responses and API failures clearly. - Verify the change against agreed sample queries and briefly explain the fix. We’ll provide repository access, setup instructions, and examples that reproduce the problem. Please confirm the cause and proposed approach before making changes. Budget: Fixed price for this specific issue. If the investigation reveals a larger problem, we’ll discuss the scope before proceeding. This is part of a longer-term project, and there may be follow-up work as we add more integrations. For now, we’re looking for a developer who can resolve this issue carefully and leave the code easy to maintain. When applying, briefly describe your experience debugging AI agent tool calls and what you would check first.