ProjectSend critical flaw actively exploited in the wild, experts warn

Researchers warn that a critical security flaw in ProjectSend open-source file-sharing application may be under active exploitation.

VulnCheck researchers warn that ProjectSend vulnerability CVE-2024-11680 (CVSS score: 9.8) appears to have been exploited by attackers in the wild.

The vulnerability is an improper authentication issue that impacts ProjectSend versions before r1720.

Remote, unauthenticated attackers can exploit this flaw by sending crafted HTTP requests to options.php, enabling unauthorized modification of the application’s configuration. Successful exploitation allows attackers to create accounts, upload webshells, and embed malicious JavaScript.

ProjectSend is an open-source file-sharing web application. Censys indexed over 1,500 GitHub stars and more than 4,000 instances exposed online. The vulnerability was published on November 26, the patch has been publicly available since May 16, 2023. The researchers pointed out that since the patch release, multiple exploits have been published by research teams, including Synactiv, Project Discovery (Nuclei), and Rapid7 (Metasploit).

“VulnCheck noticed that public-facing ProjectSend servers had started to change their landing page titles to long, random-ish strings. Some of the “random” names have larger groupings, for example” read the advisory published by VulnCheck. “These long and random-ish names are in line with how both Nuclei and Metasploit implement their vulnerability testing logic. Both exploit tools modify the victim’s configuration file to alter the sitename (and therefore HTTP title) with a random value.”

VulnCheck experts believe that threat actors started using the exploit code released by Project Discovery and Rapid7 since September 2024.

Attackers also enabled user registration, a non-default setting, to gain post-authentication access, altering the landing page to prompt account creation.

“Given how widespread we are seeing this setting enabled, we think this is likely a bigger problem than “researchers intrusively checking for vulnerable versions.” We are likely in the “attackers installing webshells” territory (technically, the vulnerability also allows the attacker to embed malicious JavaScript, too, which could be an interesting and different attack scenario).” continues the advisory.

The webshells uploaded by attackers are stored in a predictable location (upload/files/), with filenames based on upload timestamps, username hashes, and the original file name.

Defender can identify exploitation attempts by analyzing server access logs and checking for direct access to upload/files/, which bypasses intended download endpoints.

“Using the Shodan data, we found approximately 1% were using the patched version (r1750). 55% are using r1605 (released October 2022), 44% are using an unnamed release (released April 2023), and the remaining 1% are using the patched r1750.” concludes the advisory. “Given the timeline, evidence of exploitation, and lack of patch adoption, we assume that exploitation is likely widespread. And if not now, then in the near future considering the abysmal patching rates.”

Follow me on Twitter: @securityaffairs and Facebook and Mastodon

Pierluigi Paganini

(SecurityAffairs – hacking, CVE-2024-11680)