CVE-2025-8848
📋 TL;DR
This vulnerability allows HTML injection via crafted Accept-Language headers in librechat version 0.7.9. When exploited, attackers can inject arbitrary HTML into the <html lang=""> tag, potentially enabling cross-site scripting (XSS) attacks. All users running the vulnerable version are affected.
💻 Affected Systems
- danny-avila/librechat
📦 What is this software?
Librechat by Librechat
⚠️ Risk & Real-World Impact
Worst Case
Full XSS exploitation leading to session hijacking, credential theft, or complete account compromise through malicious script execution in victim browsers.
Likely Case
Limited XSS attacks that could steal session cookies or perform actions on behalf of authenticated users.
If Mitigated
Minimal impact with proper input validation and output encoding preventing HTML injection.
🎯 Exploit Status
Exploitation requires authenticated user access and specific header manipulation.
🛠️ Fix & Mitigation
✅ Official Fix
Patch Version: 0.7.10 or later
Vendor Advisory: https://huntr.com/bounties/a05ebc1f-882a-4adc-b178-d3cefa4b730e
Restart Required: No
Instructions:
1. Update librechat to version 0.7.10 or later. 2. Verify the update by checking the package version. 3. No restart required for typical deployments.
🔧 Temporary Workarounds
Input Validation Filter
allImplement server-side validation to sanitize Accept-Language header values before processing.
Implement regex filter: /^[a-zA-Z\-_,;=.*\s]+$/ for Accept-Language headers
🧯 If You Can't Patch
- Implement WAF rules to block malicious Accept-Language header patterns.
- Disable or restrict user access to vulnerable endpoints until patched.
🔍 How to Verify
Check if Vulnerable:
Check if running librechat version 0.7.9 and test with crafted Accept-Language header containing HTML payload.Check Version:
npm list librechat or check package.json for versionVerify Fix Applied:
After updating, test with same payload; HTML should be properly encoded or rejected.📡 Detection & Monitoring
Log Indicators:
- Unusual Accept-Language header patterns containing HTML tags or script elements in HTTP logs.
Network Indicators:
- HTTP requests with Accept-Language headers containing suspicious characters like <, >, or script tags.
SIEM Query:
source="web_logs" AND http_header="Accept-Language" AND (http_header CONTAINS "<" OR http_header CONTAINS ">" OR http_header CONTAINS "script")