CVE-2024-3101

7.2 HIGH

📋 TL;DR

This vulnerability in mintplex-labs/anything-llm allows attackers to disable Multi-User Mode via improper input validation, enabling them to create new admin accounts without passwords and gain unauthorized administrative access. All users running vulnerable versions of anything-llm are affected.

💻 Affected Systems

Products:
  • mintplex-labs/anything-llm
Versions: Versions before commit 52fac844221a9b951d08ceb93c4c014e9397b1f2
Operating Systems: All platforms running anything-llm
Default Config Vulnerable: ⚠️ Yes
Notes: Affects all deployments with the vulnerable code, regardless of configuration.

📦 What is this software?

⚠️ Risk & Real-World Impact

🔴

Worst Case

Complete system compromise with administrative access, allowing data theft, system manipulation, and further privilege escalation.

🟠

Likely Case

Unauthorized administrative access leading to data exposure, configuration changes, and potential lateral movement within the system.

🟢

If Mitigated

Limited impact with proper network segmentation and access controls, though administrative functions could still be compromised.

🌐 Internet-Facing: HIGH - Directly exploitable via HTTP requests without authentication when exposed to internet.
🏢 Internal Only: MEDIUM - Requires network access but can be exploited by internal threat actors or compromised internal systems.

🎯 Exploit Status

Public PoC: ⚠️ Yes
Weaponized: CONFIRMED
Unauthenticated Exploit: ⚠️ Yes
Complexity: LOW

Simple curl command with specific parameter can trigger the vulnerability as described in references.

🛠️ Fix & Mitigation

✅ Official Fix

Patch Version: Commit 52fac844221a9b951d08ceb93c4c014e9397b1f2 and later

Vendor Advisory: https://github.com/mintplex-labs/anything-llm/commit/52fac844221a9b951d08ceb93c4c014e9397b1f2

Restart Required: Yes

Instructions:

1. Update to latest version or apply commit 52fac844221a9b951d08ceb93c4c014e9397b1f2. 2. Restart the anything-llm service. 3. Verify Multi-User Mode cannot be disabled via unauthorized requests.

🔧 Temporary Workarounds

Network Access Restriction

linux

Restrict network access to anything-llm instance to trusted IPs only

iptables -A INPUT -p tcp --dport <anything-llm-port> -s <trusted-ip> -j ACCEPT
iptables -A INPUT -p tcp --dport <anything-llm-port> -j DROP

Reverse Proxy Validation

all

Configure reverse proxy to block requests containing 'multi_user_mode' parameter

# nginx example: location / { if ($args ~* "multi_user_mode") { return 403; } }

🧯 If You Can't Patch

  • Isolate the anything-llm instance from untrusted networks using firewall rules
  • Implement strict authentication and authorization controls at network perimeter

🔍 How to Verify

Check if Vulnerable:

Check if current version is before commit 52fac844221a9b951d08ceb93c4c014e9397b1f2 by examining git history or version metadata

Check Version:

git log --oneline -1

Verify Fix Applied:

Attempt to send curl request with 'multi_user_mode=false' parameter and verify it fails or requires proper authentication

📡 Detection & Monitoring

Log Indicators:

  • HTTP requests containing 'multi_user_mode' parameter
  • Unauthorized admin user creation events
  • Multi-User Mode disablement logs

Network Indicators:

  • HTTP POST/PUT requests with 'multi_user_mode=false' parameter
  • Unusual admin account creation traffic

SIEM Query:

http.method IN ("POST", "PUT") AND http.uri_params CONTAINS "multi_user_mode"

🔗 References

📤 Share & Export