Reported by 2 outlets
New ClosedQuorum Windows malware uses AI for attack decisions
Why this score
- Breadth
- 1.00 / 2.0 · 2 sources
- Authority
- 1.80 / 2.0 · best outlet weight
- Severity
- 0.50 / 5.0 · CVSS + content signals
- Velocity
- 0.00 / 2.0 · updates in 48h
- Pickup speed
- 0.60 / 0.6 · time to second source
- Recency
- × 0.00 temporal multiplier
- Community
- + 0.00 / 1.5 · reddit signal
- Score
- 0.0 / 10 = base × recency + community
Analysis
CLOSEDQUORUM is a newly identified Windows malware implant discovered by Cisco Talos through its CAIRN research toolkit. It uses commercial AI models (DeepSeek, Qwen, Mistral, Google Gemini) as an autonomous command layer: after infection it gathers host details, sends structured prompts, applies plurality voting on the responses, and executes the winning action (credential theft, code injection, persistence, or lateral movement). The sample is a 16.4 MB 64-bit Go executable. Talos has not confirmed active exploitation, but links development artifacts to carding-related criminal forums since 2025. This is the first documented Windows implant built for AI-driven tactical C2 decisions.
What matters: malware is moving from remote-operator control to on-host AI autonomy. Defenders should track AI-integrated artifacts through CAIRN-style classification, not just traditional C2 indicators. Watch whether CLOSEDQUORUM variants appear in active campaigns and whether vendors restrict model access for malicious prompts.
Timeline
-
original reporting
New ClosedQuorum Windows malware uses AI for attack decisions https:// bleepingcomputer.com/news/security/ new-closedquorum-windows-malware-
www.bleepingcomputer.com · 21d ago · September 22nd, 2026
New ClosedQuorum Windows malware uses AI for attack decisions https:// bleepingcomputer.com/news/security/ new-closedquorum-windows-malware-
via X https://www.bleepingcomputer.com/news/security/new-closedquorum-windows-malware-uses-ai-for-attack-decisions/
-
update from bleepingcomputer.com
ClosedQuorum Windows malware uses AI for attack decisions
bleepingcomputer.com · 21d ago · September 22nd, 2026
A new Windows malware named ClosedQuorum uses Google Gemini, DeepSeek, Qwen, and Mistral AI models to autonomously determine the actions to take during post-compromise stages of an attack. The Go-based malware acts with no commands from a human operator, using reconnaissance information and a voting system to decide its next step on infected hosts. When votes are tied, DeepSeek has priority at making the final call, based on the option it considers most appropriate, followed by Qwen, Mistral, and Gemini. Cisco Talos researchers analyzing ClosedQuorum say that the models are restricted to a predefined set of decisions, including: steal — simultaneously runs LSASS credential dumping, browser (Chrome, Edge, Firefox) credential theft, and cryptocurrency-wallet (MetaMast, Exodus, Ethereum) extraction inject — generates shellcode, then uses either process hollowing or Early Bird APC injection persist — executes the malware’s persistence module move — included as a possible lateral movement decision, but the analyzed distribution build has no corresponding handler, so it cannot execute this action The stolen details are then passed to the operators via a Discord webhook, so apart from the malware delivery, the attack can be fully automated. ClosedQuorum attack model overview Source: Cisco Talos describes ClosedQuorum as the first publicly documented Windows implant to delegate tactical command-and-control (C2) decisions to a panel of AI models, commenting that this adds greater spee
-
update from cyberpress.org
CLOSEDQUORUM Windows Malware Uses AI Models for Autonomous C2 and Credential Theft
cyberpress.org · 20d ago · September 23rd, 2026
The malware was identified through Talos’ newly released CAIRN research project, which tracks artifacts associated with AI-integrated malware. Talos said it has not confirmed CLOSEDQUORUM in active attacks, but linked development artifacts to criminal-forum activity related to carding dating back to 2025. Unlike conventional malware that waits for commands from an attacker-controlled command-and-control (C2) server, CLOSEDQUORUM can independently decide what to do after it infects a system. It gathers details about the compromised Windows host, sends structured prompts to AI providers, receives a decision, and executes the selected malicious activity. The sample is a 16.4 MB, 64-bit Windows executable written in Go. Each model receives host context and must choose from a limited list of actions, including credential theft, code injection, persistence, or lateral movement. The malware collects the providers’ responses and applies plurality voting. If votes are tied, the malware gives preference to DeepSeek, followe