AI Wiki · Mediocre Security
IncidentIncidents

PaperCut AI agent campaign

Aug-Sep 2026 campaign in which an attacker used AI to build PaperCut exploits and hundreds of agents to compromise 440 servers.

What happened

PaperCut published an advisory on Aug 27 2026 for two flaws in its self-hosted NG/MF print management software, CVE-2026-81578 and CVE-2026-82078, which chain into pre-authentication remote code execution; they had already been exploited as zero-days. On Aug 31, according to GreyNoise, a likely Russian-speaking actor used AI to develop and test exploits in a lab with PaperCut and an Active Directory server, then ran hundreds of AI agents on OpenAI’s Codex harness with a DeepSeek model (not OpenAI models) plus public offensive tools.

GreyNoise counts at least 440 compromised instances at 395 organizations in 48 countries. The actor went from an empty workspace to its first real-victim RCE in just under four hours and to domain admin two hours later; once the campaign launched it compromised at least 11 organizations in 26 seconds. It reached domain admin at only 12 organizations, once in seven minutes at a US high school. It tried to avoid 28 countries but did not always succeed, and in one case Cloudflare’s web application firewall stopped it.

Google’s Nov 2026 threat report also notes that researchers and Huntress used LLMs to build working exploits within a day of disclosure, and that the campaign is believed to have relied on LLM-built exploits (see exploit_agent framework). The case is tracked in AI and CVEs: latent flaws, now exploitable.

Timeline

Every dated event in the wiki that involves PaperCut AI agent campaign, newest first.

Connections

PaperCut AI agent campaign and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Concept AI and CVEs: latent flaws, now exploitable

Incident exploit_agent framework

Company GreyNoise

Sources

  1. GreyNoise, “Agents Gone Wild: An AI-Orchestrated Global Campaign Against PaperCut NG/MF”, Sep 9 2026
  2. SiliconANGLE, “Google says an AI framework turned security patches into exploits with no human in the loop”, Nov 16 2026