SesameOp backdoor exploits OpenAI Assistants API for covert command-and-control
Microsoft researchers discovered SesameOp, a backdoor that exploits OpenAI's Assistants API to establish covert command-and-control channels for malware. The malware enables persistent unauthorized access, command execution, and data exfiltration by disguising malicious activity as legitimate API traffic.
What happened
In July 2025, Microsoft Detection and Response Team researchers discovered a backdoor named SesameOp that exploits OpenAI's Assistants API. The malware uses the API as a covert command-and-control channel, enabling persistent access to compromised devices without relying on traditional methods.
The backdoor abuses the Assistants API as a storage and relay mechanism to fetch commands, which the malware then executes, while also exfiltrating data via encrypted API traffic. Microsoft researchers concluded this does not represent a vulnerability but rather a misuse of built-in API capabilities.
Reported impact
The threat actor used the SesameOp backdoor to conduct long-term espionage operations, managing compromised devices undetected and harvesting data for espionage-type purposes.
Index analysis
By leveraging OpenAI's infrastructure, the cybercriminals masked their malicious activities within legitimate API traffic, making detection more difficult than traditional command-and-control methods.
Governance controls
Microsoft urged organizations to audit and review firewalls, implement tamper protection, and deploy endpoint detection tools to mitigate threats from the SesameOp backdoor.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.