Handing off to a specialist on the AI Agents track. A handoff transfers the goal, the known facts, and the constraints. It does not transfer the whole messy transcript unless the specialist needs it. The sender stops acting after the handoff.
This lesson assumes you already worked through When several agents help.
The idea in practice
Use a structured packet: goal, facts, what was already tried, and the question for the specialist. The specialist's reply comes back as an observation.
A concrete check
goal = {
'track': 'AI Agents',
'lesson': 'Handing off to a specialist',
}
checks = [
'input available at decision time',
'score matches the real decision',
'failure case written down',
]
print(goal['lesson'])
for item in checks:
print('-', item)
Run the sketch locally if you have Python. The printout is a reminder of the checks, not a trained model. Replace the strings with the real inputs from your own example before you treat it as a design.
What usually goes wrong
Two agents both calling refund tools after a sloppy handoff duplicate the action. When this happens, stop adding parameters or tools. Fix the check, the data, or the permission, then run the same example again.
What to write down
- The input you are allowed to use at decision time.
- The output and the score or pass rule.
- One failure you will test on purpose.
- What you will not claim the system can do.
Practice
Write a handoff packet from a billing agent to a shipping agent for one delayed order.
Self-check
- Say Handing off to a specialist in one sentence that mentions an input and an output.
- Name the failure mode in this lesson and the check that would catch it.
Done when: you can explain this lesson without the page open, and you have a written failure case.