SentinelOne Inc. and Snyk Ltd. have introduced new tools to secure artificial intelligence agents. SentinelOne’s Prompt AI Agent Security identifies and protects MCP servers used by AI agents, preventing data leaks and unauthorized access. The tool is paired with Prompt AI Red Teaming, which simulates cyberattacks to find vulnerabilities in AI applications.
SentinelOne’s Singularity platform now includes an AI-native data pipeline, acquired through a $225 million startup purchase. This pipeline organizes telemetry data before uploading it, reducing unnecessary data by up to 80%.
A new AI workload security tool, Prompt Security On-Premises, allows companies to store telemetry locally rather than in the cloud, making it suitable for air-gapped environments. Ana Pinczuk, president of product and technology at SentinelOne, highlighted the need for balancing AI security with data control.
Snyk’s Evo AI-SPM automatically inventories AI assets and scans for risks. It powers three AI agents that help developers ensure the security of third-party components and test applications for vulnerabilities.