Threat actors are actively exploiting CVE-2026-33017, a critical unauthenticated remote code execution (RCE) vulnerability in Langflow, to deploy a Monero cryptocurrency miner. The campaign targets exposed artificial intelligence application endpoints, leveraging the flaw's severity to compromise systems remotely.
Assigned a CVSS score of 9.3, the vulnerability enables unauthenticated code execution without user interaction. The attacks appear systematic, with adversaries scanning for vulnerable Langflow instances and deploying the miner payload to hijack system resources for illicit cryptocurrency mining.
The attack vector exploits the RCE flaw directly, bypassing authentication mechanisms. Indicators of compromise include unauthorized outbound connections to mining pools and unexpected CPU usage spikes, though specific payload details remain limited in public reporting.
No official patch or timeline for a fix has been confirmed for CVE-2026-33017 at this time. Organizations running exposed Langflow instances are advised to restrict network access, implement authentication, and monitor for suspicious activity pending an official update.
Attribution for the campaign remains unknown. The incident highlights the growing intersection of AI infrastructure risks and cryptocurrency mining operations, as attackers continue targeting widely adopted frameworks for financial gain.