A tool call times out. What should the agent get back?
Why A is the answer
A clear error that it can act on
Readable errors let the agent retry or change its plan.
In detail
- A readable error lets the agent retry or change its plan.
- Readable errors help a model adapt.
Use case
An illustrative scenario, not a real incident.
A tool times out, and the agent gets a clear message, so it tries another path.
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{"What does the agent receive?"}
Q -->|"A clear error"| A0["It can adapt"]
Q -->|"Silence"| A1["It may hang"]Key point: Errors should be actionable.
Why the other options are wrong
- B. An empty success result and no other information
- That tells the agent it worked.
- C. A restart of the entire conversation from the beginning
- That wastes progress.
- D. An answer that was made up by the tool's wrapper
- That creates false data.

