A tool returns a 200-page document, and the agent loses track of its goal. What fixes this?
Why C is the answer
Return a short, relevant result
Tool output uses context, so return only what is needed.
In detail
- Tool output uses context, so return only what is needed.
- Tool output uses context.
Use case
An illustrative scenario, not a real incident.
A tool returns a whole manual, and the agent loses its goal. A short result fixes it.
See it in different ways
sequenceDiagram participant M as Model participant A as Application participant T as Tool M->>A: Tool name and arguments A->>A: Validate arguments A->>T: Call T-->>A: Result or error A-->>M: Result as context
sequenceDiagram participant M as Model participant A as Application participant T as Tool M->>A: Tool name and arguments A->>A: Validate against the schema A->>T: Run, with an idempotency key T-->>A: Short result or clear error A-->>M: Result as context
flowchart TD
Q{"How long is the result?"}
Q -->|"Short and relevant"| A0["Keeps the goal in view"]
Q -->|"Huge"| A1["Fills the context"]Key point: Return only what is needed.
Why the other options are wrong
- A. Add more tools for the model to choose between
- More tools add confusion.
- B. Repeat the whole document in the next model call
- That repeats the problem.
- D. Increase the sampling temperature for the next step
- Temperature does not help.