SentinelOne, an AI security provider, announced an upcoming integration of its Prompt Security offering with Amazon Bedrock AgentCore, aiming to extend existing security controls into agentic workflows for enterprises deploying AI on Amazon Web Services (AWS).
Amazon Bedrock AgentCore serves as a platform designed for building, connecting, and optimizing AI agents at scale. Its policy engine is engineered to enforce real-time, deterministic controls at the gateway across all agent traffic, encompassing communications between agents and tools, agents and large language models (LLMs), and agent-to-agent interactions.
Through this integration, SentinelOne’s Prompt Security will deliver detection signals that cover prompt injection detection, personally identifiable information (PII) exposure, tool-use validation, LLM response monitoring, and data leakage prevention. These signals will feed directly into AgentCore’s policy engine. AgentCore will then act on these signals at the gateway, operating outside the agent’s reasoning loop, which allows enterprise teams to apply SentinelOne’s detection signals alongside deterministic, automated enforcement consistently across all agent activity. This approach ensures consistent allow-or-deny decisions on every agent action, designed to support the adoption of agentic AI at enterprise scale with existing SentinelOne security controls.
Melissa K. Smith, SVP of Global Strategic Partnerships and Initiatives at SentinelOne, stated that agentic AI has become foundational for enterprise operations. She added that organizations building and running AI agents on AWS require the same security controls used across their other infrastructure. This integration is intended to apply those security guardrails at the AgentCore gateway, where all agent traffic is evaluated, to provide enterprises with confidence in deploying AI agents with their existing SentinelOne policies.
The integration is scheduled to become available in Amazon Bedrock AgentCore later this year. Customers with existing Prompt Security licenses will be able to apply them through a Bring Your Own License path. Further information is available on the SentinelOne website.
SentinelOne, headquartered in Mountain View, California, specializes in AI security, offering a platform that provides unified coverage across endpoints, identity, cloud, and AI. The company’s Autonomous Security Intelligence technology is designed to stop attacks at machine speed. SentinelOne states it protects nearly one-fifth of the Fortune 500 and hundreds of Global 2000 enterprises globally.
