agent_tool_useTier 1 · 70% confidence

ai-agents-agent-tool-use-structured-chat-zero-shot-react-description-agent--87f9daa8

agent: ai_agents

When does this happen?

IF STRUCTURED_CHAT_ZERO_SHOT_REACT_DESCRIPTION agent finishes the chain without actually calling any tool, only outputting a promise to use it.

How others solved it

THEN Switch to 'chat-conversational-react-description' agent type. If the tool requires multiple inputs, refactor it to accept a single string argument (e.g., a JSON object) and parse it internally. This avoids the multi-input tool issue that causes the agent to skip execution.

class ComparablesTool(BaseTool):
    name = "get_comparables"
    description = "... Action input should always be an object inside a string with mentioned parameters"
    def _run(self, filters_string):
        filters = json.loads(filters_string)
        # use params like filters.get('latitude')
        ...

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics