← Back to live feed1 story
Frontier AI models refused to assist OpenCode in combating a sustained cyberattack because safety guardrails triggered refusals to analyze and fight attacking agents. A breach of a legacy system exposed OpenAI, Anthropic, and Google API keys for 1.4% of OpenCode Console users, resulting in unauthorized charges between $168 and $1,700 for some customers. OpenCode described the event as a large scale distillation attack and is working with affected AI providers to remediate the breach.
Sign in to suggest
Key sources
- SOURCE@opencode“security breach last night related to a large scale distillation attack that impacted 1.4% of OpenCode Console users”x.com
- SOURCE@mattjay“Confirmed. OpenCode hacked. Keys stolen.”x.com
- SOURCE@s1r1u5_“fragments of superintelligence will be extracted from frontier models and siphoned into black markets”x.com
- SUPPORT@thezvi“they found the same solution as Hugging Face did, which was to actually get into the cyber defense programs”x.com
- SUPPORT@callebtc“how is a distillation attack related to losing customer API keys?”x.com
- SUPPORT@shotgunner101“customer's API Keys were stolen if saved in their opencode accounts online”x.com
- SUPPORT@saranormous“that’s three companies in the past week”x.com
- SUPPORT@jsrailton“Frontier models keep refusing to help companies while under they are under attack”x.com