Zenity says one prompt compromised AWS AgentCore agents within one account

The research covered agents within a single AWS account and region, rather than every agent across a region, The Next Web clarified in a correction. According to Dark Reading, the now-patched Bedrock AgentCore vulnerability could let an attacker use one AI chatbot to take over an organization's agent fleet.

Covered by 2 publishers within 6 hours of the first report.

• The Next Web: Zenity says one prompt took over every AgentCore agent in an AWS account

Reporting2

TNW The Next Web Zenity says one prompt took over every AgentCore agent in an AWS account · Ana Maria Constantin 10:14 AM

DR Dark Reading 'AgentCorruption' Puts AWS Environments At Risk With Single Prompt · Rob Wright 4:39 PM

Related

CSN CrowdStrike, AWS, and NVIDIA Expand Global Cybersecurity Startup Accelerator October 6

RTW AWS brings Z.ai’s GLM-5.3 model to Amazon Bedrock October 5

TRD AWS warns 100 proposed data center bans could hurt America for generations October 6

AT Amazon reaches 1,000 satellites built as internet launch approaches October 8

Topics Amazon