An agent audit trail is the ordered record of what an AI agent did: the actions it took, the systems it touched, the answers that came back and what authorized each one.
A log line is whatever the application chose to write. An audit trail is ordered, attributed and made of what actually passed: the request that went out, the answer that came back, who sent which. If the entries can be rewritten in place, it is a diary, not a trail.
"Who did this" cannot live in the message, because anyone can write a field. FlurryPORT records which key signed each capture, so attribution is a property of the record rather than a claim inside it.
A trail is worth what it outlives. Captures age out with your plan’s window; a question you will ask in three weeks needs a window longer than three weeks. Pick the window before you need the history, not after.
The glossary is the canonical index of every FlurryPORT term these pages use.