When should an assistant refuse, escalate, or continue with restrictions?

Instruction: Explain how you decide among refusal, restricted continuation, and escalation.

Context: Choosing refusal, escalation, or a permitted restricted path.

Updated

Example Answer

I’d decide from the request, the user’s permissions, and the consequences of the action. If the request is prohibited, I’d refuse that part and offer a permitted alternative. If it might be allowed but needs authorization or specialist judgment, I’d pause and send it to the right reviewer. If I can safely complete a smaller part, I’d do that with the limit made clear. For example, I could draft a refund explanation, but issuing the refund would still require the account checks and approval the business expects. A confident model response would not change those requirements.

Concrete example

Keep the drafted explanation separate from the actual refund operation. Record whether the request is prohibited, awaiting review, or allowed within a defined scope.

Limit to explain

Escalation is not permission to proceed while waiting. The execution layer must keep the restricted action blocked, and approval must stay within the reviewed scope.

Follow-up to practice

What should happen when no authorized reviewer is available?

Your preparation path

Work through these questions in order. Read the answer aloud, then explain it in your own words.

1. Start with the foundations

Build the vocabulary and explain the core decisions.

  1. Why is prompt injection different from ordinary bad input? Free sample
  2. How do you separate policy enforcement from model style instructions? Free sample
  3. What information should never flow from untrusted content straight into tool execution? Free sample

2. Apply it to a real workflow

Practice diagnosis, validation, and everyday tradeoffs.

  1. How do you decide where to place guardrails in a multi-step agent workflow? Free sample
  2. A summarization workflow passes unsafe content from retrieval into a downstream action. How would you fix it? Member answer
  3. A reviewer approves a high-risk action because the model framed it confidently. How would you reduce that risk? Member answer

3. Prepare for senior discussions

Explain failure boundaries, recovery, and production choices.

  1. A multi-agent system routes around one guardrail because another agent has broader permissions. How would you fix it? Member answer
  2. OAuth discovery metadata points your agent at an internal URL. How would you prevent SSRF? Member answer
  3. Why can a localhost MCP server still be exposed to attacks from a website? Member answer

Related Questions