CVE-2026-0763: GPT Academic RCE Vulnerability

CVE-2026-0763 Overview

CVE-2026-0763 is a critical insecure deserialization vulnerability affecting GPT Academic, an open-source AI-powered academic assistant platform. The vulnerability exists in the run_in_subprocess_wrapper_func function and allows remote attackers to execute arbitrary code on affected installations without requiring authentication. Due to insufficient validation of user-supplied data, attackers can exploit this flaw to achieve code execution with root privileges.

Critical Impact

This vulnerability enables unauthenticated remote code execution with root-level privileges, allowing complete system compromise of affected GPT Academic installations.

Affected Products

• GPT Academic (all versions prior to patch)
• Systems running vulnerable run_in_subprocess_wrapper_func implementations
• Deployments exposed to network access without additional access controls

Discovery Timeline

• 2026-01-23 - CVE CVE-2026-0763 published to NVD
• 2026-01-26 - Last updated in NVD database

Technical Details for CVE-2026-0763

Vulnerability Analysis

This vulnerability is classified as CWE-502: Deserialization of Untrusted Data. The flaw resides within the run_in_subprocess_wrapper_func function of GPT Academic, where user-supplied data is deserialized without proper validation. Deserialization vulnerabilities of this nature are particularly dangerous because they allow attackers to craft malicious serialized objects that, when processed by the vulnerable function, execute arbitrary code.

The vulnerability requires no authentication to exploit and can be triggered remotely over the network. Successful exploitation results in code execution within the context of the root user, providing attackers with complete control over the compromised system. This includes the ability to read and modify sensitive data, install persistent backdoors, and pivot to other systems on the network.

Root Cause

The root cause of this vulnerability is the lack of proper validation and sanitization of user-supplied data before it is passed to deserialization routines within the run_in_subprocess_wrapper_func function. The function accepts serialized data from external sources and processes it without verifying the integrity or trustworthiness of the input, enabling object injection attacks.

Attack Vector

The attack vector for CVE-2026-0763 is network-based, requiring no user interaction or authentication. An attacker can exploit this vulnerability by sending specially crafted serialized payloads to the vulnerable run_in_subprocess_wrapper_func endpoint. When the application deserializes this malicious data, it instantiates attacker-controlled objects that can execute arbitrary system commands.

The exploitation process typically involves:

1. Identifying an exposed GPT Academic instance
2. Crafting a malicious serialized payload containing code execution gadgets
3. Sending the payload to the vulnerable function endpoint
4. Achieving remote code execution as root upon deserialization

For detailed technical information regarding this vulnerability, refer to the Zero Day Initiative Advisory ZDI-26-029.

Detection Methods for CVE-2026-0763

Indicators of Compromise

• Unexpected network connections to GPT Academic services from external IP addresses
• Anomalous subprocess spawning or execution patterns on systems running GPT Academic
• Suspicious serialized data payloads in application logs or network traffic captures
• Unauthorized processes running with root privileges linked to the GPT Academic application

Detection Strategies

• Monitor network traffic for unusual serialized object patterns targeting GPT Academic endpoints
• Implement application-level logging to capture all calls to run_in_subprocess_wrapper_func and analyze input data
• Deploy intrusion detection rules to identify known deserialization attack payloads
• Use endpoint detection and response (EDR) solutions like SentinelOne to detect post-exploitation behavior

Monitoring Recommendations

• Enable verbose logging for GPT Academic application components
• Configure network monitoring to alert on unexpected inbound connections to GPT Academic services
• Implement file integrity monitoring on critical system files and application directories
• Review process execution logs for suspicious child processes spawned by the application

How to Mitigate CVE-2026-0763

Immediate Actions Required

• Restrict network access to GPT Academic installations using firewall rules or network segmentation
• Implement authentication requirements for all external-facing GPT Academic endpoints
• Consider temporarily taking vulnerable instances offline until patches are available
• Deploy web application firewalls (WAF) with rules to filter serialized data payloads

Patch Information

Monitor the GPT Academic project and the Zero Day Initiative Advisory ZDI-26-029 for official patch releases. Apply security updates immediately upon availability. Organizations should establish a process to track upstream security advisories for this and other dependencies.

Workarounds

• Deploy network-level access controls to restrict which hosts can communicate with GPT Academic
• Implement input validation at the application gateway layer to reject suspicious serialized payloads
• Run GPT Academic in a sandboxed or containerized environment with reduced privileges
• Consider disabling or restricting access to the run_in_subprocess_wrapper_func functionality if not required

# Example: Restrict network access to GPT Academic using iptables
# Allow access only from trusted internal network
iptables -A INPUT -p tcp --dport 8080 -s 10.0.0.0/8 -j ACCEPT
iptables -A INPUT -p tcp --dport 8080 -j DROP