What happened
Microsoft Incident Response researchers discovered a new backdoor called SesameOp that uses OpenAIâs Assistants API for command-and-control communications in compromised environments. The backdoor was found during an investigation into a sophisticated, months-long security incident where attackers maintained persistent access. SesameOp leverages internal web shells and compromised Microsoft Visual Studio utilities to execute commands remotely.
Why it matters
- The attack demonstrates a novel misuse of legitimate AI services for stealthy, long-term espionage operations.
- SesameOpâs use of the OpenAI API as a covert channel makes detection more difficult compared to traditional methods.
- The incident highlights ongoing risks from advanced, persistent threats targeting enterprise environments.
Key facts
- 1. SesameOp was discovered in July 2025 during a Microsoft Incident Response investigation.
- 2. Attackers maintained access within the targeted environment for several months before detection.
- 3. The backdoor uses .NET AppDomainManager injection to compromise Microsoft Visual Studio utilities.
