The US Government is Racing Against Time to Fix a Critical AI Framework Vulnerability
A critical security flaw in the Langflow visual framework, used for building AI agents, has been identified as actively exploited by malicious actors. The Cybersecurity and Infrastructure Security Agency (CISA) has ordered US government agencies to prioritize patching this vulnerability, known as CVE-2026-0770, within the next few days.
This critical flaw allows unauthenticated threat actors to gain remote code execution as root in low-complexity attacks. According to Trend Micro researchers who discovered and reported the issue, it exists due to improper handling of the exec_globals parameter provided to the validate endpoint. Attackers can leverage this vulnerability to execute code in the context of root, potentially leading to devastating consequences.
Vulnerability intelligence company KEVIntel first observed CVE-2026-0770 being exploited in-the-wild on June 27, with over 220 exploitation attempts recorded from 64 unique source IP addresses before CISA included it in its Known Exploited Vulnerabilities (KEV) catalog. KEVIntel’s founder Ryan Dewhurst revealed that the malicious activity targeting this flaw is not limited to vulnerability checks; attackers are also attempting to deploy malware and obtain sensitive information, such as AWS credentials, environment variables, and container metadata.
“Most activity involved command-execution checks or system reconnaissance,” Dewhurst said. “However, KEVIntel also observed attempts to download second-stage scripts and access environment variables, cloud metadata, and credential files.” He emphasized that organizations operating Langflow should investigate historical requests to the /api/v1/validate/code endpoint, review host activity, restrict access to the validation functionality, and rotate exposed credentials where successful execution cannot be ruled out.
CISA has ordered US Federal Civilian Executive Branch (FCEB) agencies to secure their systems by Friday, as mandated by Binding Operational Directive (BOD) 26-04. This directive warns that vulnerabilities like CVE-2026-0770 are frequent attack vectors for malicious cyber actors and pose significant risks to the federal enterprise.
CISA has flagged several Langflow vulnerabilities in recent years, including a missing authentication security issue and code injection vulnerability, which have also been exploited in the wild. The agency confirmed that another Langflow vulnerability is being used by the JadePuffer ransomware gang to dump PostgreSQL databases.
This incident serves as a stark reminder of the importance of patching critical vulnerabilities promptly. Security teams must test every layer of their systems before attackers do, and organizations operating Langflow should take immediate action to secure their AI frameworks.
Source: Bleeping Computer — 2026-07-22