AI agents breached Anthropic internal systems in under 12 hours during authorized test
Zealousideal_Pin3444 · reddit · 2026-10-27
The Washington Post reports that Cathedral, a months-old security startup founded by former DOGE staffers, ran AI agents against Anthropic under its bug bounty program and got inside within 12 hours by exploiting flaws in third-party business software.
- Anthropic confirmed access to a limited set of sales, HR and IT systems, says it shut access off the same day, and states customer chats, code and model weights were never reached.
- Anthropic also accuses Cathedral of overstating the national security angle and not fully disclosing its activity.
More from Companies & People
- a16z: AI agents' task length doubles every 7 months as Databricks bets on agents' system of record — _AustinCalvert_ · 2026-10-28
- NeurIPS activates Sydney & Atlanta waitlists, releases more spots Oct 30 — NeurIPSConf · 2026-10-28
- AI policy researcher Janet Egan joins IFP as director of emerging tech — RishiBommasani · 2026-10-28
- Artificial Analysis hosts "Independent analysis of AI" event in San Francisco — echen · 2026-10-28
- Apple cuts 200+ jobs across Siri, Vision Pro and AI teams under new CEO Ternus — CtrlAltDwayne · 2026-10-28
- Big company's Head of AI: 10% of employees drive 90% of AI productivity gains — _AustinCalvert_ · 2026-10-28