Each sweep runs one probe at a time. Clearing a probe takes a reason, and the third blocker is the last one you get.
Picked up where you left it.
Name the risk you are knowingly letting through. That line is the one nobody writes, and it is why people trust the review.
Paste this into the PR, the doc thread, or the incident record.
Which probes actually catch things in your hands, and which ones you have never once used. The second list is the more interesting one.
Run but never produced a finding. Either your systems are clean there, or you are not looking hard.
Paste the JSON record from an agent-run sweep (the review-sweep playbook), or the whole memo with the record at the end.