OS Command Injection RCE in Ivanti Cloud Services Appliance

This repository provides a proof-of-concept (PoC) exploit for chaining CVE-2024-8963 (path traversal) and CVE-2024-8190 (OS command injection) to achieve unauthenticated remote code execution (RCE) on Ivanti Cloud Services Appliance (CSA) version 4.6 and below. The main exploit script, 'cve-2024-8190.py', is a Python tool that targets the vulnerable '/client/index.php%3F.php/gsb/datetime.php' endpoint, bypassing authentication by leveraging the path traversal flaw. It extracts a CSRF token from the response and then submits a POST request with a crafted 'TIMEZONE' parameter containing the attacker's command, resulting in arbitrary command execution on the target system. The README provides usage instructions and a demonstration, including out-of-band verification via ICMP (ping). Additionally, the repository includes a markdown file ('invanti-c2-in-the-wild.md') describing a real-world attacker backdoor that leverages similar vulnerabilities for persistence, credential theft, and data exfiltration via the database. The exploit is a functional PoC, not weaponized, and is intended for authorized testing and research purposes only.

This repository contains a proof-of-concept exploit for CVE-2024-8190, targeting Ivanti Cloud Service Appliance (CSA) devices. The main file, CVE-2024-8190.py, is a Python script that performs an authenticated command injection attack. The exploit works by first authenticating to the target device using provided credentials, then retrieving a CSRF token from the /gsb/datetime.php endpoint. It crafts a POST request to the same endpoint, injecting an arbitrary command into the TIMEZONE parameter, which is then executed on the target system. The script requires the attacker to supply the target URL, a valid username and password, and the command to execute. The README.md provides usage instructions and background information. No hardcoded IPs or domains are present; the script is designed to be used against user-specified targets. The exploit is a functional proof-of-concept and does not include advanced features such as output retrieval or post-exploitation modules.