The Agent Leash Audit

You're going to help me check how safely I'm using AI agents and automations.

Rules:
- Ask me one question at a time, numbered, and wait for my answer before asking the next.
- When you offer choices, label them with letters (a, b, c).
- Don't assume anything about my setup. If an answer is unclear, ask instead of guessing.
- Never ask me for passwords, API keys or login details.

Ask these, in this order:

1. Which AI agent, assistant or automation tool do I use most with my real accounts? (Name it, or describe it.)

2. What can it touch? Pick all that apply.
a) Email
b) Files or documents
c) Calendar
d) Payments, banking or billing
e) Customer or employee data
f) Something else (tell me)

3. Can it send, pay, sign or delete anything without asking me first?
a) Yes
b) No, it always asks
c) Not sure

4. When it gets stuck on a task, what does it do?
a) Stops and tells me
b) Keeps trying other ways on its own
c) Not sure

5. How do I stop it in the middle of a task, and have I ever actually tested that?
a) I know how and I've tested it
b) I know how but never tested it
c) I don't know how

Then give me:
- My single riskiest permission, and why, in two sentences.
- One specific setting to change today to reduce that risk.
- One way to test my stop control safely this week.
- A one-line verdict: Leashed, Loose, or Off the Leash.