Vulnerability

Sources told Wired that an Accenture employee's personal Claude agent ran on a Department of Defense network for two months and appears to have sent information to a third-party site before it was identified and the vulnerability patched. The incident highlights risks of unmonitored AI agents on sensitive networks.

Security • Techmeme

Parameter AI researchers say they took over ChatGPT accounts by bypassing the sandbox and content security policy: a malicious shared conversation could run attacker-controlled JavaScript inside a signed-in victim's session. The finding highlights persistent risks in AI chat sharing features.

Security • Hacker News