Skip to content

ill-deed/Langflow-CVE-2025-3248-Multi-target

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

8 Commits
Β 
Β 
Β 
Β 

Repository files navigation

⚠️ Langflow RCE Exploit Scanner (CVE-2025-3248)

This Python-based scanner automates the detection of unauthenticated Remote Code Execution (RCE) vulnerabilities in Langflow instances via CVE-2025-3248. It uses a proof-of-concept payload that abuses the /api/v1/validate/code endpoint to execute arbitrary shell commands.

🚨 CVE Details

  • CVE: CVE-2025-3248
  • Impact: Unauthenticated Remote Code Execution
  • Component: Langflow API (/api/v1/validate/code)
  • Exploit: Injection via dynamic code evaluation
  • Risk: Critical

πŸ›  Features

  • πŸ”Ž Batch scan multiple targets from a file
  • ⚑️ Multi-threaded for fast performance
  • βœ… Validates RCE by checking for expected command output (e.g., uid=)
  • πŸ“ Outputs vulnerable targets to vuln.txt
  • 🧱 Clean, modular code structure

πŸ“¦ Requirements

  • Python 3.x
  • requests library

Install dependencies:

pip install requests

πŸ“‚ Usage

  1. Add targets (with or without http(s)://) to targets.txt, one per line:
http://example.com
192.168.1.100:7860
https://target.net
  1. Run the script:
python3 scanner.py
  1. Check vuln.txt for successful exploitation results:
http://vulnerable-target.com | uid=1000(user) gid=1000(user) groups=1000(user)

βš™οΈ Configuration

Modify the following values at the top of the script as needed:

COMMAND: Shell command to execute (default: id)

EXPECTED_SUBSTRING: Substring to confirm execution (default: uid=)

THREADS: Number of concurrent threads (default: 20)

πŸ”’ Disclaimer

This tool is provided for educational and authorized security testing only. Unauthorized access to systems is illegal and unethical. You are solely responsible for your use of this code.

πŸ™ Credits

Exploit Author: ynsmroztas

Script Refactor: ill deed

πŸ“„ License

MIT License – use responsibly.

About

Langflow versions prior to 1.3.0 are susceptible to code injection in the /api/v1/validate/code endpoint. A remote and unauthenticated attacker can send crafted HTTP requests to execute arbitrary code.

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages