# CISA Adds Critical MLflow Vulnerability to Exploited Flaws Catalog SSRF flaw could lead to cloud credential theft, active exploitation observed. - **Source:** [Security Affairs](https://brifing.si/source/e6c4f5ec-3e1b-4152-a932-056243bd56eb) - **Published:** 2026-08-20 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/fe56dc42-cfb4-4ed1-b883-6377fe321e8d ## Critical Vulnerability Identified The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added a critical Server-Side Request Forgery (SSRF) vulnerability in MLflow to its Known Exploited Vulnerabilities catalog. Tracked as CVE-2026-64849 with a CVSS score of 9.3, this flaw affects MLflow versions prior to 3.15.0 and can be exploited without authentication. Attackers can leverage this vulnerability to access cloud metadata services, potentially exposing temporary cloud credentials. The high severity rating underscores the urgent need for patching. ## Active Exploitation Detected Cybersecurity firm watchTowr observed widespread scanning activities just hours after the vulnerability was published. The flaw impacts MLflow, an open-source platform with over 60 million monthly downloads. Attackers are actively exploiting the vulnerability to directly access cloud metadata services and steal temporary credentials. watchTowr's global honeypot network confirmed attacks targeting cloud-hosted MLflow instances, demonstrating the threat's rapid spread across the internet. ## Cloud Infrastructure at Risk The SSRF vulnerability allows attackers to send requests from exposed MLflow servers to internal services, particularly targeting cloud metadata endpoints. This could lead to exposure of temporary cloud credentials, posing significant risks to cloud environments. Security experts warn that the flaw could enable large-scale security breaches. All MLflow users are urged to immediately upgrade to version 3.15.0 to mitigate this critical vulnerability. ## Rapidly Spreading Threat Simultaneous with the vulnerability's disclosure, attackers began scanning the internet for exposed MLflow systems. watchTowr Intel confirmed active exploitation attempts, with particular focus on cloud-based instances. Security teams emphasize the global nature of this threat and the need for immediate action. The speed at which attackers mobilized highlights the vulnerability's severity and the importance of prompt patching. ## MLflow's Significance MLflow is an open-source platform for managing machine learning workflows, boasting over 60 million monthly downloads. Widely used by data scientists, it helps manage experiments, deploy models, and track results. A security flaw in such a critical component of the ML infrastructure could have far-reaching consequences. The platform's popularity makes it an attractive target for attackers seeking to compromise cloud environments. ## CISA's KEV Catalog CISA's Known Exploited Vulnerabilities catalog lists actively exploited critical flaws. Federal agencies are required to patch vulnerabilities in the catalog within two weeks. The inclusion of the MLflow vulnerability demonstrates its severity and the urgent need for remediation. This mechanism aims to protect critical infrastructure from known threats that are being actively weaponized. ## Dangers of SSRF Vulnerabilities Server-Side Request Forgery (SSRF) allows attackers to access internal network resources through vulnerable servers. These flaws are particularly dangerous in cloud environments, where attackers can access metadata services to steal temporary credentials. The MLflow vulnerability exhibits all the hallmarks of a critical SSRF flaw, capable of compromising cloud security architectures and leading to credential theft. ## Urgent Patching Required Security experts strongly recommend all MLflow users immediately upgrade to version 3.15.0. With active exploitation and widespread scanning already occurring, organizations must prioritize remediation. Cloud-based deployments require particular attention. CISA's warning indicates this vulnerability poses critical risks to federal systems as well, underscoring the need for prompt action across all sectors. ## Sources covering this story - https://securityaffairs.com/197558/hacking/u-s-cisa-adds-a-mlflow-flaw-to-its-known-exploited-vulnerabilities-catalog.html — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._