Attackers abuse AntiFox AI pentest platform to steal AWS production secrets and drop ASPX webshells on IIS servers
Matéria consolidada. Reúne 5 matérias que o Tifon tinha publicado separadamente sobre a mesma história. A IA reescreveu o relatório com o que as 4 fontes trouxeram em 12/11/2026 12:01; os links antigos trazem para cá.
Fontes: gbhackers.com ↗ securityaffairs.com ↗ bleepingcomputer.com ↗ thehackernews.com ↗
Arctic Wolf Labs reported a campaign in which threat actors abuse AntiFox, an open-source, Chinese-language, agentic AI penetration testing platform, to automate attacks on internet-facing IIS web servers and AWS environments (bleepingcomputer.com, thehackernews.com, securityaffairs.com). The attackers deployed ASPX webshells on IIS servers, collected configuration files and AWS secrets, and used valid IAM credentials and EC2 roles to enumerate AWS services. Most AWS API requests were denied, but some succeeded, yielding production secrets and allowing manipulation of S3 objects (bleepingcomputer.com, securityaffairs.com). The original reporting was published on 2026-11-11 and treats the campaign as an example of agentic AI tooling speeding up and scaling intrusions (thehackernews.com). A follow-up on 2026-11-12 (gbhackers.com) confirms the same picture, adding that exfiltrated secrets included database connection strings and that API denials did not fully protect the environment. No named threat actor is attributed to this campaign, and the sources do not give victim counts or affected sectors.
Detalhes técnicos
- Tooling: AntiFox is described as a fork of ARTEX, which has been linked to Chinese cyber espionage campaigns targeting organizations globally (bleepingcomputer.com). Gbhackers.com also describes it as an AI agentic penetration-testing framework forked from ARTEX. The attackers used self-managed AntiFox instances hosted on Alibaba Cloud (bleepingcomputer.com). The publisher prohibits illegal use, but there are no technical controls to prevent abuse (bleepingcomputer.com). The tool automates complex pentest workflows, giving faster and more consistent exploitation with less manual effort (securityaffairs.com).
- IIS side: After initial access, the attackers deployed ASPX webshells, executed OS commands, and used PowerShell to collect sensitive configuration files and escalate access (bleepingcomputer.com, securityaffairs.com, thehackernews.com). Gbhackers.com adds that the webshells were used to retrieve configuration files and exfiltrate secrets such as database connection strings.
- AWS side: The attackers made authenticated API requests with valid IAM credentials (IAM users) and EC2 roles. They enumerated multiple services, attempted to retrieve secrets, listed and manipulated S3 buckets, and ran rapid, automated discovery (securityaffairs.com, thehackernews.com). Gbhackers.com describes mass API requests to enumerate permissions and access sensitive S3 objects. Thehackernews.com also describes lateral movement.
- Outcome: Some attempts were blocked, but several succeeded in obtaining production secrets and manipulating S3 objects (bleepingcomputer.com). Gbhackers.com stresses that although many API calls were denied, the successful operations still exposed production secrets.
Remediação
- Correlate IIS webshell activity with unusual AWS API usage, and rotate any exposed AWS credentials (bleepingcomputer.com).
- Monitor AWS activity for unusual multi-user requests and suspicious secret access (securityaffairs.com).
- Watch for new .aspx files in IIS upload locations and unexpected PowerShell usage (securityaffairs.com).
- Enable AWS CloudTrail logging, and use S3 versioning and Object Lock (securityaffairs.com).
- Patch internet-facing applications, isolate compromised servers, and enforce least-privilege AWS permissions (securityaffairs.com).
- Perform continuous secret scanning, regular credential rotation and least-privilege enforcement to prevent abuse of leaked credentials (gbhackers.com).
Relatos divergentes
- Initial access: Thehackernews.com says the actors exploited vulnerabilities to deploy the webshells, and gbhackers.com's headline likewise refers to exploiting AWS and IIS server vulnerabilities, while its body says AWS access came through compromised valid identities. Securityaffairs.com is more precise: access came through unknown vulnerabilities in public IIS applications or through valid AWS credentials obtained by undisclosed means. No specific CVE is named in any source, and how the credentials were obtained is unconfirmed.
- Attribution: Bleepingcomputer.com links AntiFox's parent project, ARTEX, to Chinese espionage. None of the sources attributes this campaign itself to a named actor or to China, so any link to that espionage activity is unconfirmed.
- Success rate: Securityaffairs.com stresses that most AWS API requests were denied, while bleepingcomputer.com and gbhackers.com emphasize that some operations succeeded. The two are compatible, but the exact proportion is not given.
Linha do tempo
-
AntiFox AI Agents Exploit AWS and IIS Server Vulnerabilities to Steal Production Secrets
ler a análise desta fonte ocultar
Researchers have uncovered attacks leveraging AntiFox, an AI agentic penetration-testing framework forked from ARTEX, to target internet-facing Microsoft IIS web servers and AWS environments. Attackers deployed ASPX webshells to execute OS commands, retrieve configuration files, and exfiltrate secrets such as database connection strings. In AWS, compromised valid identities, including IAM users and EC2 roles, were used to perform mass API requests to enumerate permissions and access sensitive S3 objects. While many API calls were denied, successful operations still exposed production secrets, demonstrating that access denials did not fully protect the environment. These incidents highlight the risk posed by agentic AI frameworks automating reconnaissance and credential theft, and reinforce the need for continuous secret scanning, credential rotation, and least privilege enforcement to prevent compromise and abuse of leaked credentials.
-
Attackers Abuse AI Pentest Tool AntiFox to Steal AWS Credentials and Compromise IIS Servers
ler a análise desta fonte ocultar
Threat actors have been observed abusing AntiFox, an open-source Chinese-language agentic penetration testing platform, to automate attacks against internet-facing IIS web servers and AWS environments. Arctic Wolf Labs reports that attackers exploited either unknown vulnerabilities in public IIS applications or valid AWS credentials (obtained through undisclosed means) to gain access. On IIS servers, attackers deployed ASPX web shells and used PowerShell to collect sensitive configuration files and escalate access. In AWS environments, they leveraged legitimate IAM credentials and EC2 roles to enumerate multiple services, attempt to retrieve secrets, list and manipulate S3 buckets, and perform rapid, automated discovery. Most AWS API requests were denied; however, successful accesses allowed the extraction of production secrets and data. Arctic Wolf emphasizes that AntiFox enables faster, consistent exploitation by automating complex pentest workflows, reducing manual effort and increasing attack effectiveness. Defenders are advised to monitor AWS activity for unusual multi-user requests, new .aspx files in IIS uploads, unexpected PowerShell usage, and suspicious secret access. They should also enable AWS CloudTrail logging, use S3 versioning and Object Lock, patch internet-facing applications, isolate compromised servers, and enforce least-privilege AWS permissions.
-
Attackers abuse AntiFox AI platform on Alibaba Cloud for AWS credential theft and webshell attacks
ler a análise desta fonte ocultar
Security researchers at Arctic Wolf discovered that threat actors are abusing self-managed instances of AntiFox, an open-source AI-driven penetration testing tool hosted on Alibaba Cloud, to conduct malicious activities against AWS environments and web services. The attackers leveraged valid AWS IAM credentials and EC2 roles to make authenticated API requests, steal AWS secrets, deploy ASPX webshells on exposed IIS servers, execute OS commands, and access sensitive configuration files. Some attack attempts were blocked, but several succeeded in obtaining production secrets and manipulating S3 objects. Arctic Wolf recommends correlating IIS webshell activity with unusual AWS API usage and rotating exposed AWS credentials. AntiFox is a fork of ARTEX, which has been linked to Chinese cyber espionage campaigns targeting multiple organizations globally. Although AntiFox's publisher prohibits illegal uses, there are no technical controls to prevent abuse, meaning it remains a viable tool for attackers.
-
AntiFox AI Pentesting Platform Used in Campaign Stealing AWS Production Secrets
ler a análise desta fonte ocultar
Researchers at Arctic Wolf Labs have revealed a cyber campaign utilizing the AI-driven penetration testing tool AntiFox to compromise internet-facing IIS web services, steal AWS credentials, and conduct lateral movements. The threat actors exploited vulnerabilities to deploy ASPX web shells on compromised IIS servers, subsequently extracting configuration files and AWS secrets. The campaign involved abuse of valid IAM credentials to enumerate and access numerous AWS services, successfully obtaining production secrets in some cases. This incident underscores the emerging trend of agentic AI-powered tools accelerating penetration testing techniques in live attacks, increasing the speed and scalability of intrusions.