ai-agents-tool-call-validation-pydantic-validationerror-when-the-llm-returns-tool-1f5318b2

THEN Check your tool function signatures: if a parameter expects a plain string but the model returns a dict (e.g., {'category': 'math'}), update the tool to unwrap the dict. For production, ensure your tool schemas clearly define expected types and consider post-processing the arguments before calling the tool.

def GetNewQuestion(category) -> str:
    if isinstance(category, dict):
        category = category.get('category', 'general')
    return f"What is your experience with {category}?"