Preliminary research

AgentCorruption: How a Single Prompt Collapsed the Entire Cloud Security Model

Tamir Ishay Sharbat and Lana Salameh · Zenity Labs · 8 October 2026

AI-collected research leads through 3 November 2026, including a November–October–September missed-article review and a targeted multilingual search alongside earlier community-source sweeps. Unranked, incomplete, not community-vetted, and subject to change.

An overview of AgentCore HTTP-tool access to own-role metadata credentials and the broad permissions that expose sibling agents, conversations, memories and connector secrets within one AWS account and region. Links four technical companions and includes AWS's expected-behavior response.

Record

Researcher

Tamir Ishay Sharbat and Lana Salameh

Published by

Zenity Labs

Date

8 October 2026

In the archive

• Collected in the 2026 preliminary research sweep

Related sources

• Initial IMDS Access
• One Role to Rule Them All
• Credential Theft from AWS Secrets Manager and More
• Weaponizing Agent Memory for Persistent Hijacking
• Credential theft demo

Tags

• prompt-injection
• ssrf
• aws
• iam
• credential-theft
• memory-poisoning
• owasp-a03-2021
• owasp-a10-2021

This page is the archive's own catalogue record. The research is the work of Tamir Ishay Sharbat and Lana Salameh, first published at the original source. Preserved copies are kept so the citation survives its host.