CVE-2025-63663
📋 TL;DR
This vulnerability allows unauthorized attackers to access other users' uploaded files through the /api/v1/conversations/*/files API in GT Edge AI Platform. It affects all users of GT Edge AI Platform versions before v2.0.10 due to improper access control mechanisms.
💻 Affected Systems
- GT Edge AI Platform
📦 What is this software?
⚠️ Risk & Real-World Impact
Worst Case
Complete compromise of all user-uploaded files including sensitive documents, credentials, and proprietary data leading to data breaches and regulatory violations.
Likely Case
Unauthorized access to confidential files belonging to other users, potentially exposing personal information, business documents, or intellectual property.
If Mitigated
Limited exposure if proper network segmentation and authentication controls are in place, but still vulnerable to authenticated attackers.
🎯 Exploit Status
Exploitation requires API access but no specific user permissions. Public proof-of-concept available in GitHub repository.
🛠️ Fix & Mitigation
✅ Official Fix
Patch Version: v2.0.10
Vendor Advisory: https://gist.github.com/p80n-sec/f3ca933480157cb4e18c387d92f4d0c2
Restart Required: Yes
Instructions:
1. Backup current configuration and data. 2. Download v2.0.10 from official vendor source. 3. Stop the GT Edge AI Platform service. 4. Apply the update following vendor documentation. 5. Restart the service and verify functionality.
🔧 Temporary Workarounds
API Endpoint Restriction
allTemporarily block or restrict access to the vulnerable /api/v1/conversations/*/files endpoint
# Using iptables (Linux)
iptables -A INPUT -p tcp --dport <api_port> -m string --string "/api/v1/conversations/" --algo bm -j DROP
# Using Windows Firewall
New-NetFirewallRule -DisplayName "Block GT Edge AI Vulnerable API" -Direction Inbound -Protocol TCP -LocalPort <api_port> -RemoteAddress Any -Action Block
Authentication Enforcement
linuxImplement additional authentication layer or API gateway with strict access controls
# Example nginx configuration
location /api/v1/conversations/ {
auth_request /auth;
proxy_pass http://backend;
}
🧯 If You Can't Patch
- Implement network segmentation to isolate the GT Edge AI Platform from untrusted networks
- Deploy a web application firewall (WAF) with rules to detect and block unauthorized file access patterns
🔍 How to Verify
Check if Vulnerable:
Test if unauthorized API requests to /api/v1/conversations/*/files return other users' file data. Check current version against vulnerable range.Check Version:
Check platform version in admin interface or run: curl -X GET http://<host>:<port>/api/v1/versionVerify Fix Applied:
After patching, verify that unauthorized requests to the API endpoint return proper access denied errors and only authenticated users can access their own files.📡 Detection & Monitoring
Log Indicators:
- Multiple failed authentication attempts followed by successful file access
- Unusual patterns of file access from single IP across multiple user sessions
- API requests to /api/v1/conversations/*/files with mismatched user IDs
Network Indicators:
- Unusual volume of requests to file API endpoints
- Requests accessing file paths with sequential or random conversation IDs
- Traffic patterns showing horizontal movement across user accounts
SIEM Query:
source="gt_edge_ai.log" AND (uri_path="/api/v1/conversations/*/files" AND user_id!=requesting_user_id)