Langflow RCE Used to Deploy Monero Miner on Exposed AI Endpoints

Colleagues, I’d like to highlight a cybersecurity update.

Langflow vulnerability CVE-2026-33017 is already being actively abused to deploy a Monero miner on exposed AI endpoints.

Attackers:
- gain unauthorised remote code execution;
- download a shell script and miner binary;
- establish persistence via cron, SSH keys, and disabling safeguards;
- spread to other SSH-accessible systems.

This shows that exposed AI infrastructure is becoming a real entry point into the corporate network.

If you have Langflow or similar AI services exposed to the internet, they should be closed, access-restricted, patched, and monitored for signs of cryptojacking.

How do you protect AI endpoints in your environment?