Research Identified 94 High-Impact Security Vulnerabilities
The findings highlight an uptick in critical security flaws and the integration of AI agents into malicious operational workflows.
Updated on Oct. 13, 2026 in Cybersecurity
Live Poll
Do you believe the rise of AI-assisted cyberattacks puts your personal data at greater risk?
Insikt Group identified 94 high-impact vulnerabilities throughout September 2026, with 59 of those classified as having a Very Critical Recorded Future Risk Score. Nearly half of the discovered flaws allowed for remote code execution.
Why it matters
The surge in identified vulnerabilities illustrates a significant expansion in the attack surface, further compounded by threat actors now leveraging automated AI agents to execute large-scale compromises.
The study tracked 94 high-impact vulnerabilities across 58 vendors, 47 of which enabled remote code execution. Of these, 14 were detected exclusively through honeypot data, while 75 now have public proof-of-concept exploits.
The players
Insikt Group
A threat intelligence unit that monitors global digital infrastructure, vulnerability disclosures, and emerging cyber-adversary tactics.
CISA
The Cybersecurity and Infrastructure Security Agency, which maintains the Known Exploited Vulnerabilities catalog as a primary reference for urgent patching.
The details
Researchers utilized a multi-vector monitoring approach, tracking the CISA Known Exploited Vulnerabilities (KEV) catalog alongside independent open-source reporting and honeypot data collection. The investigation revealed that threat actors have begun using AI agents to conduct reconnaissance and automate attacks. One campaign involved a suspected Russian-speaking actor exploiting 440 PaperCut NG/MF instances across 395 organizations, while a separate Chinese-speaking actor utilized an environment backed by Claude, Qwen, and DeepSeek.
Timeline
September 2026: Insikt Group identified 94 high-impact vulnerabilities.
The Tech Race
This vulnerability reporting aligns with the reliance on the CISA Known Exploited Vulnerabilities catalog as the industry standard for risk prioritization. It marks a shift as adversarial efforts move beyond manual targeting toward AI-automated exploitation workflows.
Organizations should cross-reference the 43 vulnerabilities identified in the CISA KEV catalog against their internal software inventories for immediate prioritization. Those running PaperCut NG/MF instances should audit system logs for the specific compromise patterns attributed to the recently observed AI-agent campaigns.
The takeaway
The rapid identification of vulnerabilities and the rise of AI-driven exploitation underscore the need for automated, intelligence-led patch management. Security teams should monitor future updates from threat intelligence providers for new proof-of-concept releases related to the 75 vulnerabilities identified here.
Further reading
For more on evolving threat landscapes and mitigation strategies, visit Cybersecurity.
Source note: This article includes information reported by Recordedfuture.
Live Poll
Do you believe the rise of AI-assisted cyberattacks puts your personal data at greater risk?
