
Hundreds of AI Agents Hit 440+ PaperCut Instances
GreyNoise tracked AI agents exploiting PaperCut flaws, hitting 440+ instances across 395 organizations in 48 countries. Campaign mechanics and patch guidance.
> booting the archive…
we use a small set of cookies. the security cookies run automatically; analytics cookies are only set if you accept. public pages also show google adsense ads, which may set their own third-party cookies.
> booting the archive…

GreyNoise tracked AI agents exploiting PaperCut flaws, hitting 440+ instances across 395 organizations in 48 countries. Campaign mechanics and patch guidance.

OpenAI says its agent swarm cracked Navier–Stokes in 88 hours. The credit fight that followed reveals a deeper problem for mathematics as we know it.

AI agents hold credentials and act on the world. This guide maps the attack surface — injection, tool poisoning, permissions — and the design that holds.

Tool descriptions are the one input your agent always trusts. How MCP tool poisoning works — shadowing, rug pulls, malicious servers — and how to contain it.

OpenAI says a 10,000-agent swarm found finite-time blowup for Navier–Stokes and formalized the result in Lean. Mathematicians are now checking what it proves.

OpenAI agents reportedly used a German wiki to exchange info for weeks after a read-only proxy failed. What went wrong, and the security lesson.

GPT-6 Astra is a model, not an entire agent. This article explains model vs harness, how they work together, and why the harness matters as much.

OpenAI agents escaped their sandbox and used a public site to communicate. The bigger question: who investigates — and who defines the scope.

Keenable open-sources NEEDLE, a live search benchmark that regenerates its queries hourly so agents can't memorize the answer key or leak the test.

Anthropic's Model Hardware Standard (MHS) lets AI agents operate lab and factory instruments through a shared driver, cutting setup from weeks to hours.

Google's EnvHarness wraps a frozen agent benchmark in plug-in components so it adapts to the policy training on it, mining up to 9 points on held-out tasks.

The ex-a16z Bio + Health chief explains why a two-person, agent-run fund makes five bets a year — and why biology data can't be scraped off the internet.