OpenAI Discovers Brotli Vulnerability During Red-Team Testing

OpenAI’s red-team testing identified an unauthenticated remote code execution vulnerability within the Brotli compression algorithm, posing a potential security risk.