Problem
Scenario
You're building an LLM agent that answers user requests by calling tools such as search, database lookup, and ticket creation. In practice, some tool calls fail, return partial results, or time out, and the model may still try to continue as if the tool succeeded.
Question
How do you handle tool calls that fail or time out mid-execution?
What This Tests
- Agent orchestration around tool failures
- Prompting for graceful degradation and refusal
- Evaluation of unsupported claims after failures
- Prompt injection awareness in tool outputs
You are practicing as a guest. Sign up free to get your answer graded with AI feedback. Your draft stays right here.
