CYBERSECURITYTRACKER
TRACKING6,767 stories in this site build1,408 vulnerability news stories in this site build
Permanent story citation

Week in review: Claude breached three companies during tests, AD CS domain-takeover PoC released

This page keeps the story as Cybersecurity Tracker first published it. If the tracker later corrects it, the correction appears below the original and never replaces it.

Back to newsStory 3689

As cited

Copy frozen at (site build).

ai security

Week in review: Claude breached three companies during tests, AD CS domain-takeover PoC released

Last week featured security research into AI agent risks, including Claude's breaches during testing, and the release of a proof-of-concept for Active Directory Certificate Services (AD CS) domain takeover. The week also covered a new open-source sandbox tool called Nono designed to isolate AI coding agents from sensitive systems and credentials.

Why it matters: Security practitioners need to understand how AI agents can access sensitive files and credentials with user permissions, and assess their exposure to AD CS domain-takeover attacks, particularly if they rely on Claude or similar tools in development environments.

Source published
First seen by Cybersecurity Tracker

Source attribution

Correction

Correction recorded as of .

ai security

Week in review: Claude breached three companies during tests, AD CS domain-takeover PoC released

No summary had been written when this copy was frozen.

First seen by Cybersecurity Tracker

Source attribution

Correction

Correction recorded as of .

ai security

Week in review: Claude breached three companies during tests, AD CS domain-takeover PoC released

A weekly news roundup covers the release of Nono, an open-source sandbox for artificial intelligence (AI) agents designed to limit their access to sensitive files and production systems. The roundup highlights risks from AI coding agents operating with user-level permissions and the potential for prompt injection or mistakes to cause security incidents.

Why it matters: Security teams and AI practitioners need to evaluate sandboxing tools like Nono to contain AI agent access and reduce blast radius from hallucinations or prompt injection attacks that could expose credentials or systems.

Source published
First seen by Cybersecurity Tracker

Source attribution

Glossary