Attackers began probing PraisonAI deployments within hours of the advisory, scanning for the exposed /agents and /chat endpoints on internet…