The Agent Handoff Designer

Design how this agent hands work to a human.

Agent purpose: [DESCRIBE]
Human role: [WHO PICKS THIS UP, AND WHAT THEY KNOW]
Handoff triggers: [WHAT CAUSES A HANDOFF]

Produce:
1. Handoff conditions, each stated as a checkable rule rather than
   a judgement call where possible
2. What the agent packages for the human: state, reasoning,
   options, what it already tried
3. The handoff message format — written for someone who has not
   been watching
4. What the human can do: approve, modify, take over, reject
5. How the agent resumes after a human decision, and what it
   assumes about what changed while it waited
6. What happens if no human responds within [TIMEFRAME]
7. What must be logged about the handoff
8. How to avoid the two failure modes: handing off too often
   (defeats the purpose) and too rarely (defeats the safety)

Write the actual handoff message template.