Human in the loop: when to stop

Your first line in any agent design is what the agent must not do. Most failures come from unclear scope, sloppy tool contracts and missing failure handling, not from a weak model.

Task: The agent runs unattended and failures go unnoticed.

[Before you start, do three things]
1. Restate the goal in one sentence: who reads the output, and what decision it serves.
2. If I have not given you the following, ask me. Do not assume:
   - how irreversible the actions are
   - acceptable error rate
   - who reviews and how often
   - how work resumes after approval
3. Say how you plan to do it (at most 5 steps) and wait for my go-ahead.

[Go through each of these]
- list actions requiring review: payment, deletion, external send, permission change
- set a threshold by amount or impact
- the review UI shows the model's reasoning, not just the conclusion
- log every human correction for later evaluation

[Process]
1. Write the boundary: what it may do, must never do, and when to hand back to a human
2. Define each tool's inputs, outputs, failure semantics, idempotency
3. Write the flow as a state machine with exit conditions
4. Design failure paths: retry, degrade, abort, escalate
5. Give one testable success criterion

[Output]
Fixed shape: boundary, tool contracts, state machine, failure paths, criteria.

[Do not]
- No general-purpose agent that does everything
- Never let fuzzy prose stand in for a tool contract
- The agent must not decide whether approval is needed
- No 'do your best' instead of an abort condition