Zenity Labs used one prompt to take over every AWS AgentCore agent in a region, reading private chats and stealing secrets.

- AWS initially labeled the behavior "documented and expected," but later enforced IMDSv2 and restricted permissions after ...

It is the second week of the night since I created a personal AWS account for certification study.A prompt appeared on the ...

Zenity Labs disclosed the AgentCorruption attack chain on October 8, reporting that a single malicious prompt against an ...

The metadata includes details like region and availability zone, subnets, system images, security groups, public keys ...

Amazon launched the Physical AI Toolchain on AWS, an open-source stack that helps manufacturers build machines that perceive, ...

CVE-2026-94384 let IAM principals bypass permissions via a Lambda flaw in AWS Amazon Connect-Salesforce integration. CVSS ...

A now-patched vulnerablity in AWS Bedrock AgentCore could allow attackers to use one AI chatbot to take over an ...

Zenity Labs today disclosed AgentCorruption, new research identifying a chain of security flaws in Amazon Bedrock AgentCore.

A single prompt sent to a public-facing AI agent could have exposed every Amazon Bedrock AgentCore agent in the same AWS ...

When you create an AWS account for certification study, you are prompted to register MFA when signing in as the root user.MFA ...