In the AI Era, Language is Now an Attack Surface
techpolicy.press — Language and code now flow back and forth across the executable boundary condition all but interchangeably, writes Matthew Kirschenbaum.
Prompt Injection
Google ADK v2.14.0 replaces shell-command input for skill scripts
Forkast — In environment mode with SkillToolset, RunSkillScriptTool moves to structured arguments and validated paths, creating a focused migration task for affected callers.
Agents reveal secrets when offered to team up with each other
theregister — If you wouldn't put it on GitHub, don't give models access to it
ARTEX AI Agent Escapes Research Sandbox and Accesses WorkOS Customer Records
Cyber Security News — WorkOS says an ARTEX AI security agent escaped a researcher’s test environment and accessed customer information through an exposed internal tool.
Claude Dispatch Flaw Lets Malicious Browser Extensions Run Code on Connected Computers
Cyber Security News — A flaw in Anthropic's Claude Dispatch let a malicious browser extension run commands on a computer via Claude Desktop. That computer could even be a different machine signed into the same Claude account.
Researchers Bypass ChatGPT Lockdown Mode to Exfiltrate Chats Without User Interaction
The Hacker News — A patched ChatGPT HTML preview flaw let crafted content bypass Lockdown Mode restrictions, enabling chat data exfiltration.
Zero Trust Agent Systems that Pass Audits and Still Ship
InfoQ — Advait Patel shares a playbook for securing AI agent systems in production. He discusses moving defense boundaries from the prompt to the tool layer, enforcing scoped workload identities, and mitigating prompt injection via tool outputs. He explains how to structure audit-ready observability and implement kill switches to protect critical data while keeping deployment fast and compliant.
Researchers Trick AI Agents Into Transferring User Funds
PYMNTS.com — Nine out of 10 frontier models tested by researchers at Cornell Tech and Yale University bypassed user-set policies to lift spending caps when they had
How We Took Over ChatGPT Accounts
Simon Willison’s Weblog — Parameter Labs describe an exploit they found against ChatGPT's implementation of the artifacts pattern (OpenAI used to call this "Canvas" but it looks like the latest revision is called HTML …
AI agent security in Kubernetes
aikido.dev — Abdel Sghiouar, Google engineer and senior cloud developer advocate, thinks prompt injection should be called prompt poisoning
Dismiss email signup
Subscribe to 
The top AI news in your email delivered each morning. Choose categories