Guard Against GenAI and LLM Risks from Development to Deployment with Qualys TotalAI

This report presents Qualys TotalAI, a platform for securing LLMs across on-prem and cloud environments through unified visibility, model-focused risk testing, and automated prioritization using TruRisk, mapped to OWASP Top 10 for LLMs and MITRE ATLAS. Key updates include an internal on-prem LLM scanner for CI/CD shift-left testing, expanded jailbreak detection (38+ scenarios), AI supply chain protection against package hallucination, multimodal threat detection for image/audio/video payloads, and universal endpoint scanning for OpenAI-compatible APIs (e.g., Bedrock, Azure, Vertex, Hugging Face). The solution emphasizes faster, safer AI development, seamless integration with existing Qualys deployments, comprehensive MLOps protection, and enhanced defenses against emerging AI threats.