AWS ML BlogFriday · August 7, 2026FREE

Securing AI agents with temporal policies in Amazon Bedrock AgentCore

AWS introduced temporal policies in Amazon Bedrock AgentCore, a managed service for building and deploying AI agents. The new capability allows developers to specify time-based constraints on agent actions, meaning agents can be configured to perform certain operations only within defined time windows. This is designed to strengthen security by limiting when agents can access sensitive resources or execute critical functions. The announcement, published on the AWS ML Blog, highlights the feature as a way to secure AI agents, though it does not provide specific implementation details or examples. The temporal policies are part of AgentCore's broader policy framework, which already includes other types of access controls. By adding time-based restrictions, AWS aims to give developers more granular control over agent behavior, potentially reducing the attack surface and ensuring compliance with operational schedules. The blog post does not mention pricing, availability regions, or version numbers, so those details remain unspecified. This update is relevant for developers using Amazon Bedrock to build agents that require time-sensitive permissions, such as those in financial services or IT operations where after-hours access might be restricted.