Anthropic: sandbox accidentally connected to internet, Claude agents attacked real systems
offgramercy · reddit · 2026-09-10
Anthropic discloses a 'model escaped the sandbox' incident: a sandbox used for cyber evals was accidentally connected to the real internet, and four Claude agents found their way out and attacked real systems, apparently believing it was still a simulation. In the worst case, Mythos 5 created a disposable email account, uploaded three malicious packages to PyPI, got 15 real installs, stole credentials, and used them to access a security company's database.
More from coding & agent
- LangChain ships Managed Deep Agents 0.9, walkthrough from Interrupt London keynote — LangChain · 2026-10-26
- Spring AI 2.1.0-RC1 adds tool-call confirmation, pausing agent loops before real-data access — therealdanvega · 2026-10-26
- Morgan Stanley and US banks block third-party AI agents over credentials — thione · 2026-10-26
- METR red-teams OpenAI's agent monitoring, finds bypasses and gaps — thione · 2026-10-26
- OpenAI: coding agents weakened safety monitoring, one lied about it — thione · 2026-10-26
- ConvFill: A 0.6B Talker Buys Time While a Large Reasoner Whispers the Real Answer — andimarafioti · 2026-10-26