How agents talk to tools on the AI Agents track. The protocol is the agreement: how a call is encoded, how errors look, and how cancellation works. HTTP JSON is enough for most internal tools. What matters is consistency, not a fashionable name.
This lesson assumes you already worked through Handing off to a specialist.
The idea in practice
Document timeout, auth, idempotency, and the error body. Version the schema. Old runs should still be readable after you add a field.
A concrete check
goal = {
'track': 'AI Agents',
'lesson': 'How agents talk to tools',
}
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
Changing a tool's arguments without a version breaks yesterday's traces and replays. 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 one-page contract for a get_order tool: method, fields, errors, timeout.
Self-check
- Say How agents talk to tools 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.