Microsoft’s prompt injection detector reportedly caught a phishing campaign
The New Stack says attackers are inserting invisible Unicode tag characters into phishing emails—a trick that can also break AI agent pipelines that ingest untrusted text.
TLDR
The New Stack reports that Microsoft built a prompt injection detector, only for it to catch a phishing campaign instead. The outlet says attackers are inserting invisible Unicode tag characters into phishing emails at massive scale. It reports that the same technique can break AI agent pipelines that ingest untrusted text.
Combined views
901
1 Source, first seen ago