Skip to content
Horizon3 CTO Says AI-Enabled Attackers Exploit Old Unpatched Flaws at Scale

Horizon3 CTO Says AI-Enabled Attackers Exploit Old Unpatched Flaws at Scale

First seen 21 Nov 2026, 21:46 UTC • •

Article Content

Browse articles
ThreatCluster AI
ThreatCluster •November 21, 2026 at 21:46 UTC
  • •Pillitiere claims AI-enabled attackers succeed mainly by exploiting known, unpatched, internet-facing systems rather than novel techniques.
  • •The reported initial access for the "Jade Puffer" actor was an externally exposed Jenkins CI/CD server exploited via a 2025 vulnerability.
  • •Pillitiere says NVD vulnerability additions are running at almost three times the prior weekly rate.
  • •Threat actors reportedly adopted the open-source pentesting harness "Artex"; Pillitiere says the FBI issued a flash warning about it.
  • •Defenders are advised to focus on fundamentals such as asset management, layered defense, and MFA, alongside autonomous pentesting to find gaps first.

Horizon3.ai co-founder and CTO Anthony Pillitiere argues that AI-enabled threat actors are succeeding with known, unpatched vulnerabilities rather than new exploits, according to remarks on the MTS podcast. He cites a first reported agentic threat actor, referred to as "Jade Puffer," which he says he learned of in July and which gained initial access through an externally exposed Jenkins CI/CD server by exploiting a 2025 vulnerability. Pillitiere says vulnerabilities added to the National Vulnerability Database have been arriving at almost three times the prior weekly rate, and that organizations are not patching fast enough. He describes the economics of agentic attackers as "unlimited cyber interns" where one success among many failed attempts is enough. He points to the open-source pentesting harness "Artex," which he says threat actors adopted and which he says the FBI has flagged in a flash warning. Pillitiere reads the recent wave of infrastructure outages as organizations paying down technical debt rather than as attacks. The articles present these claims as Pillitiere's assessment; no specific CVE identifiers, IOCs, or independent confirmation of the Jade Puffer campaign are provided.

Start a free Starter trial for enhanced analysis

Ask AI about this cluster

Updated 21h ago How this analysis works

Timeline

Recent
Pillitiere reports Jade Puffer details reached him
Pillitiere says he learned of the Jade Puffer actor in July, referencing a 2025 vulnerability.
Finance.Biggo
Date unknown
Jade Puffer agentic threat actor initial access via Jenkins
Pillitiere says the actor gained entry through an externally exposed Jenkins CI/CD server by exploiting a 2025 vulnerability.
Finance.Biggo

More articles in this cluster (2)

Following this threat?

Track Anthropic in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.

Free account · no card needed

Common questions

Is my Jenkins server exposed to the internet?
The articles say the Jade Puffer actor entered through an externally exposed Jenkins CI/CD server, and Pillitiere argues such systems should not face the public internet. Checking external exposure of CI/CD systems is the relevant step.
Which vulnerability did the attackers use?
The articles identify only a vulnerability from 2025 and do not give a CVE identifier. No specific patch or advisory is named.
What should defenders prioritize?
Pillitiere recommends fundamentals such as asset management, layered defense, and multi-factor authentication, plus faster patching of internet-facing systems. These are his recommendations, not an official advisory.