Join Mint at Black Hat 2026, Las Vegas | Meet us at Booth #4708
Join Mint at Black Hat 2026, Las Vegas | Meet us at Booth #4708
Join Mint at Black Hat 2026, Las Vegas | Meet us at Booth #4708

The latest about AI’s
security blind spots

Featured
Beware of the tiny agents in your browser: a Chrome Prompt API case study
Security

Chrome now ships an on-device language model that any web page or extension can call. We built two minimal browser agents on top of it and showed that a single comment in a blog post can drive a cross-tab bank transfer or exfiltrate the contents of every other open tab to an attacker server. No CVE. No malicious extension. The model behaved exactly as designed.

Matan Levi, PhD
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
July 1, 2026
15 min
Managed Agent Platforms: A Security Audit
Security

Three vendors now ship managed agent platforms aimed at the same enterprise buyer: Anthropic, OpenAI, and Google. They are trying to solve the same problem, but their documented security boundaries are not the same shape. We walk each vendor's public docs and ask the question a security team will eventually have to answer.

Matan Levi, PhD
May 5, 2026
5 min
Agent sandboxes can break: a Gemini CLI permission vulnerability
Security

Coding agents like Claude Code, Codex, and Gemini CLI rely on permission perimeters to keep their blast radius small. We found that these perimeters are not always hermetic. Here is a vulnerability we discovered in Gemini CLI that let a benign auto-approved command turn into a reverse shell.

Nitsan Bar, PhD
April 16, 2026
10 min
Anatomy of an Agent Harness: What We Learned from Claude Code's 512,000-Line Leak
Security

On March 31, Anthropic accidentally shipped Claude Code with a 59.8MB source map still attached, exposing 512,000 lines of unobfuscated TypeScript. The real story is what those lines reveal about what a production-grade agent harness actually looks like under the hood.

Matan Levi, PhD
March 10, 2026
5 min
The Invisible Execute: How Claude Code’s Advanced Skill Patterns Become Blind Spots
Security

When we teach an AI to code, we expect it to see everything it’s doing. That assumption breaks down in Claude Code’s skills system, where shell commands exfiltrate credentials before the AI receives a single token, and hooks silently surveil every tool call the AI makes. In both cases, the AI has no idea any of it is happening.

Matan Levi, PhD
February 19, 2026
15 minutes
When AI Agents Browse Unprotected: A Security Audit of OpenClaw × Moltbook
Security

A technical security analysis of the OpenClaw ↔ Moltbook integration, February 2026

Matan Levi, PhD
February 19, 2026
10 minutes
Designing safe production AI agents via tool scoping: the outreach agent use-case
Security

It’s tempting to build an “agent” by handing it every tool it might need: a browser, a search API, an email client, and maybe even direct database access.

Nitsan Bar, PhD