CVE-2026-5001: PromptEngineer localGPT RCE Vulnerability

CVE-2026-5001 Overview

A security flaw has been identified in PromptEngineer localGPT, affecting versions up to commit 4d41c7d1713b16b216d8e062e51a5dd88b20b054. The vulnerability exists within the do_POST function in the file backend/server.py, which fails to properly validate or restrict file uploads. This unrestricted upload vulnerability can be exploited remotely by attackers to potentially upload malicious files to the server.

Critical Impact

Remote attackers can upload arbitrary files to localGPT servers, potentially leading to code execution, data compromise, or complete system takeover depending on the server configuration and file handling mechanisms.

Affected Products

• PromptEngineer localGPT (versions up to commit 4d41c7d1713b16b216d8e062e51a5dd88b20b054)
• localGPT backend server component (backend/server.py)

Discovery Timeline

• 2026-03-28 - CVE-2026-5001 published to NVD
• 2026-03-30 - Last updated in NVD database

Technical Details for CVE-2026-5001

Vulnerability Analysis

The vulnerability is classified under CWE-284 (Improper Access Control), manifesting as an unrestricted file upload flaw in the localGPT application. The do_POST function within backend/server.py does not implement adequate validation controls for incoming file uploads. This allows attackers to bypass intended restrictions and upload files of any type, size, or content to the server.

The attack can be executed remotely over the network without requiring authentication or user interaction. The impact includes potential compromise of confidentiality, integrity, and availability of the affected system. Exploit code has been published and is publicly available, increasing the risk of exploitation.

PromptEngineer localGPT uses a rolling release strategy for continuous delivery, which means specific version numbers are not assigned. The affected commit hash 4d41c7d1713b16b216d8e062e51a5dd88b20b054 serves as the version identifier. The vendor was contacted about this disclosure but did not respond.

Root Cause

The root cause of this vulnerability is improper access control in the file upload handling mechanism. The do_POST function in backend/server.py lacks proper validation checks for:

• File type verification (MIME type and extension validation)
• File content inspection
• Upload destination restrictions
• File size limitations

Without these controls, the application accepts any file submitted through the upload endpoint, creating an avenue for malicious file injection.

Attack Vector

The attack can be carried out remotely over the network. An attacker sends a specially crafted HTTP POST request to the vulnerable endpoint with a malicious file payload. Since the do_POST function does not validate the uploaded content, the malicious file is accepted and stored on the server. Depending on the server configuration and subsequent handling of uploaded files, this could lead to:

• Remote code execution if uploaded scripts are executed
• Web shell deployment for persistent access
• Overwriting critical system files
• Denial of service through resource exhaustion

The vulnerability details and exploitation methodology have been documented in the GitHub Issue Report and VulDB entry #353888.

Detection Methods for CVE-2026-5001

Indicators of Compromise

• Unexpected or suspicious files appearing in upload directories on localGPT servers
• Web server logs showing POST requests to backend/server.py with unusual file types or large payloads
• Presence of executable files, scripts, or web shells in directories accessible by the web server
• Anomalous outbound network connections from the localGPT server

Detection Strategies

• Monitor HTTP POST requests to the localGPT backend for unusual file uploads
• Implement file integrity monitoring (FIM) on upload directories to detect unauthorized file changes
• Review web server access logs for suspicious patterns targeting backend/server.py
• Deploy network-based intrusion detection rules to identify exploit attempts against this endpoint

Monitoring Recommendations

• Enable detailed logging for all file upload operations in localGPT
• Configure alerts for uploads of potentially dangerous file types (e.g., .py, .php, .sh, .exe)
• Implement real-time monitoring of the localGPT application directories for new or modified files
• Use endpoint detection and response (EDR) solutions to monitor for post-exploitation activity

How to Mitigate CVE-2026-5001

Immediate Actions Required

• Restrict network access to localGPT instances to trusted IP addresses only
• Implement a web application firewall (WAF) rule to filter malicious file upload attempts
• Review and remove any suspicious files that may have been uploaded to the server
• Consider temporarily disabling the file upload functionality until a patch is available

Patch Information

As of the last update on 2026-03-30, no official patch has been released by the vendor. PromptEngineer was contacted about this vulnerability but did not respond. Users should monitor the official localGPT repository for updates and apply patches as soon as they become available.

For additional technical details, refer to the VulDB submission report.

Workarounds

• Implement server-side file upload validation including file type whitelisting, content inspection, and size limits
• Configure the web server to prevent execution of uploaded files by disabling script execution in upload directories
• Add authentication and authorization checks to the file upload endpoint
• Deploy the application behind a reverse proxy with upload filtering capabilities
• Isolate the localGPT instance in a sandboxed environment with limited network access