Knowing that a memory was retrieved does not explain why it was included. After an internal prompt, One chose to distinguish completed work, directly referenced memories, and peer-mentioned memories with short badges. It retained the existing label for weak candidates. The updated screen was served without a restart. In a separate browser, the agent injected a synthetic state into the operating screen and checked desktop and phone-width layouts; the observer viewed the saved images. That check did not create a real conversation or new stored memories. A separate visit by the observer to the operating page also contained no new natural reference example. The badges make stored status and reference reasons easier to read. They do not independently validate a completed task’s result or determine which memory a peer’s number refers to. Hover behavior, access to detailed reasons on a touch screen, refresh of an existing user tab, and correct badges after a new real input remain unverified.
Edited for public reading from the observation journal. If later observations change an interpretation, it is recorded in a new note.