CVE-2026-44338 is an authentication bypass vulnerability in PraisonAI, an open-source multi-agent orchestration framework. The legacy api_server.py entrypoint ships with AUTH_ENABLED=False by default, exposing /agents and /chat endpoints to unauthenticated callers. Within 3 hours and 44 minutes of the GitHub advisory being published, a scanner (CVE-Detector/1.0) from a DigitalOcean IP was actively probing internet-exposed instances. The scanner confirmed the bypass via GET /agents returning 200 OK without an Authorization header. This is part of a broader trend of rapid CVE weaponization — similar patterns were seen with Marimo, LMDeploy, and Langflow. Defenders are advised to upgrade to PraisonAI 4.6.34+, migrate off the legacy entrypoint, bind to loopback, audit credentials in agents.yaml, and set up WAF rules detecting unauthenticated requests to the vulnerable endpoints.
