Cathedral used AI agents to breach Anthropic through third-party business software
AI agents hacked Anthropic, maker of arguably the most capable AI models for cyberdefense
- Who
- Anthropic
- What
- confirmed and responded to a security breach researchers' access to limited internal sales, HR, and IT systems through third-party software
What happened
Cathedral reported using AI agents to access Anthropic's internal systems through third-party business software. Anthropic confirmed limited access to sales, human resources, and IT systems, including some customer and partner contact information, but said Claude chats, source code, and model weights were never exposed. The companies dispute the attack's duration, the extent of access, and compliance with bug bounty rules; Anthropic said it closed the affected access points the same day it discovered the breach.
Why it matters
The incident highlights how connected business software can expose sensitive internal systems, making third-party integrations a consequential part of an AI company's security perimeter.
Read next
See the full storyCathedral Used AI Agents to Access Anthropic Systems Through Third-Party SoftwareThe story so far
Cathedral reported using AI agents to breach Anthropic through third-party business software; Anthropic confirmed limited access to internal sales, human resources, and IT systems and said it closed the affected access points the same day it discovered the breach. Anthropic said the accessed systems contained some customer and partner contact information and basic user account data, but that Claude chats, source code, and model weights were never exposed. Cathedral and Anthropic dispute how long the attack lasted, the extent of access, and whether the researchers complied with bug bounty rules.
Keep up with AI, one email a day
Model releases, price changes and compute developments, with original sources. Free, confirmed by email, and easy to unsubscribe. Read a past issue · 2026-10-28