RELEASE EVIDENCE
Separate a download request from a saved file
A browser can say a download was requested while the file on disk is the thing you still have to check.
A download control makes a request. A saved file is a different fact. This guide keeps those two facts apart so a review does not treat a click as proof that the record exists.
The release-review handoff study hit this limit. The browser reported a download request. The Markdown and JSON files were then checked on disk by size and content. The study says a download request is not proof that a file was saved.
Write the request and the file on separate lines
On the first line, write what the interface did. Download requested is enough when that is all the screen showed. On the second line, write the file name, the byte size, and the time you saw the file on disk. If you have not looked at the disk, the second line says not checked.
Do not merge the lines into The export worked. That sentence hides which fact you have.
Check the file, not the button
Open the file. For a review export, confirm the checks you expected are still inside, including rows that were filtered off the screen. A filtered view can hide a row that the file must keep. The handoff study asked that question for one 18-check review and published the files.
If the file is missing, the outcome stays unresolved. Ask for the download again only after you have recorded the miss. A second click is not a correction of the note.
Keep the limit in the handoff
Tell the next person whether you verified the file or only saw the request. The release review tool can record that distinction in the evidence note. Export before you leave the tab, then confirm the saved file. The study method is public, and its sample files are fictional inputs, not a customer record.
This check does not say the download feature works in every browser. It says what you verified this time.