---
title: "AI agents breached Anthropic's internal systems via third-party software"
url: https://www.parallelquant.com/posts/ai-agents-breached-anthropic-s-internal-systems-via-third-party-software-0c1555
source_name: "The Decoder"
source_url: https://the-decoder.com/ai-agents-hacked-anthropic-maker-of-arguably-the-most-capable-ai-models-for-cyberdefense/
published: 2026-10-28T10:56:24.000Z
topics: ["security", "llms"]
publisher: "Parallel Quant"
---

# AI agents breached Anthropic's internal systems via third-party software

*2026-10-28 · Source: [The Decoder](https://the-decoder.com/ai-agents-hacked-anthropic-maker-of-arguably-the-most-capable-ai-models-for-cyberdefense/)*

Security startup Cathedral used AI agents to break into Anthropic's internal systems through third-party software, as part of a bug bounty program. Anthropic confirmed the access covered sales, HR and IT data, but disputes how deep the breach went.

**Why it matters:** Autonomous agents are now finding and chaining supply-chain weaknesses against well-defended targets, which shifts the economics of offense. Even a lab that markets strong cyberdefense models was exposed through vendor dependencies, a reminder that third-party software is often the weakest link.

**Topics:** security, llms

---
Read the original: https://the-decoder.com/ai-agents-hacked-anthropic-maker-of-arguably-the-most-capable-ai-models-for-cyberdefense/
Canonical: https://www.parallelquant.com/posts/ai-agents-breached-anthropic-s-internal-systems-via-third-party-software-0c1555
Published by Parallel Quant — https://www.parallelquant.com