Recently, when posting statuses that contain links, we’ve noticed an increase in false-positive “Status contains malware” errors. We treat those errors as permanent failures, so they aren’t retried automatically by our software.
In this case, the errors are almost all incorrect. The Google safe browsing diagnostic indicates that the URL is not listed as suspicious or hosting malware. In each case the statuses were resubmitted (with the same URL) successfully shortly after the failure.
The errors seem to cluster around time periods an hour or two long. Today’s was around 2pm Pacific time, but I saw a similar cluster of occurrences on October 30 and October 31. Could this be due to an intermittent outage between Twitter and the service(s) it uses to determine if a URL is safe?
EDIT 13 November at 10am: Clarifying that statuses work when resubmitted with the same URL.