CVE-2026-0772: Langflow Deserialization RCE Vulnerability

CVE-2026-0772 Overview

CVE-2026-0772 is a remote code execution vulnerability affecting Langflow, stemming from insecure deserialization of untrusted data in the disk cache service. This vulnerability allows authenticated remote attackers to execute arbitrary code on affected Langflow installations by exploiting improper validation of user-supplied data during deserialization operations.

The flaw was identified by the Zero Day Initiative and tracked as ZDI-CAN-27919. When successfully exploited, an attacker can execute code in the context of the service account, potentially compromising the entire system running Langflow.

Critical Impact

Authenticated attackers can achieve remote code execution through deserialization of malicious data in Langflow's disk cache service, enabling full system compromise under the service account context.

Affected Products

• Langflow (specific versions not disclosed in advisory)

Discovery Timeline

• 2026-01-23 - CVE CVE-2026-0772 published to NVD
• 2026-01-26 - Last updated in NVD database

Technical Details for CVE-2026-0772

Vulnerability Analysis

This vulnerability (CWE-502: Deserialization of Untrusted Data) exists within Langflow's disk cache service. The disk cache component fails to adequately validate user-supplied data before deserializing it, creating an opportunity for attackers to inject malicious serialized objects.

Deserialization vulnerabilities occur when an application processes serialized data from untrusted sources without proper validation. In Python-based applications like Langflow, this typically involves the use of serialization libraries such as pickle that can execute arbitrary code during the deserialization process.

The attack requires authentication, which provides some mitigation against opportunistic attacks. However, once authenticated, an attacker can craft malicious payloads that, when deserialized by the disk cache service, result in arbitrary code execution.

Root Cause

The root cause of this vulnerability is the lack of proper validation of user-supplied data before deserialization in the disk cache service. The application trusts serialized data without verifying its integrity or origin, allowing attackers to inject malicious serialized objects that execute code upon deserialization.

This is a common weakness in applications that use serialization for caching or data persistence without implementing secure deserialization practices such as input validation, type checking, or using safer serialization formats.

Attack Vector

The attack vector is network-based, requiring the attacker to have valid authentication credentials to the Langflow application. Once authenticated, the attacker can submit specially crafted serialized data to the disk cache service.

The exploitation process involves creating a malicious serialized object containing code execution payloads. When the disk cache service deserializes this object, the embedded code executes in the context of the service account, giving the attacker control over the affected system.

Since no verified proof-of-concept code is available, the specific exploitation technique involves crafting malicious serialized objects targeting the disk cache service. For detailed technical information, refer to the Zero Day Initiative Advisory ZDI-26-038.

Detection Methods for CVE-2026-0772

Indicators of Compromise

• Unusual process spawning from the Langflow service account
• Unexpected network connections originating from the Langflow server
• Anomalous disk cache operations or unusual serialized data patterns in cache directories
• Suspicious authentication attempts followed by cache-related API calls

Detection Strategies

• Monitor Langflow application logs for unusual deserialization errors or exceptions
• Implement application-layer monitoring to detect malformed or suspicious serialized data payloads
• Deploy network intrusion detection signatures for known deserialization attack patterns
• Use endpoint detection and response (EDR) solutions to identify post-exploitation behavior

Monitoring Recommendations

• Enable verbose logging for the disk cache service to capture deserialization events
• Configure SIEM alerts for suspicious activity patterns associated with the Langflow service account
• Monitor for child process creation from the Langflow process that deviates from normal operations
• Implement file integrity monitoring on Langflow installation directories

How to Mitigate CVE-2026-0772

Immediate Actions Required

• Review and restrict authentication access to Langflow installations
• Implement network segmentation to limit exposure of Langflow services
• Monitor for exploitation attempts using the detection strategies outlined above
• Consider temporarily disabling the disk cache feature if operationally feasible

Patch Information

Consult the Zero Day Initiative Advisory ZDI-26-038 for the latest patch information and vendor remediation guidance. Apply vendor-supplied patches as soon as they become available to address this vulnerability.

Workarounds

• Restrict network access to Langflow instances using firewall rules and access control lists
• Implement strong authentication controls and limit user accounts with access to the application
• Deploy web application firewalls (WAF) configured to inspect and block suspicious serialized data payloads
• Consider using network isolation or containerization to limit the impact of potential compromise

# Configuration example - Restrict network access to Langflow
# Add firewall rules to limit access to trusted IP addresses only
iptables -A INPUT -p tcp --dport 7860 -s trusted_ip_range -j ACCEPT
iptables -A INPUT -p tcp --dport 7860 -j DROP