Threat actors are leveraging a recently disclosed security flaw impacting Ivanti Connect Secure, Policy Secure, and ZTA gateways to deploy a backdoor codenamed DSLog on susceptible devices.

That's according to findings from Orange Cyberdefense, which said it observed the exploitation of CVE-2024-21893 within hours of the public release of the proof-the-concept (PoC) code.

CVE-2024-21893, which was disclosed by Ivanti late last month alongside CVE-2024-21888, refers to a server-side request forgery (SSRF) vulnerability in the SAML module that, if successfully exploited, could permit access to otherwise restricted resources sans any authentication.

The Utah-based company has since acknowledged that the flaw has limited targeted attacks, although the exact scale of the compromises is unclear.

Then, last week, the Shadowserver Foundation revealed a surge in exploitation attempts targeting the vulnerability originating from over 170 unique IP addresses, shortly after both Rapid7 and AssetNote shared additional technical specifics.

Orange Cyberdefense's latest analysis shows that compromises have been detected as early as February 3, with the attack targeting an unnamed customer to inject a backdoor that grants persistent remote access.

"The backdoor is inserted into an existing Perl file called 'DSLog.pm,'" the company said, highlighting an ongoing pattern in which existing legitimate components – in this case, a logging module – are modified to add the malicious code.

DSLog, the implant, comes fitted with its own tricks to hamper analysis and detection, including embedding a unique hash per appliance, thereby making it impossible to use the hash to contact the same backdoor on another device.

The same hash value is supplied by the attackers to the User-Agent header field in an HTTP request to the appliance to allow the malware to extract the command to be executed from a query parameter called "cdi." The decoded instruction is then run as the root user.

"The web shell does not return status/code when trying to contact it," Orange Cyberdefense said. "There is no known way to detect it directly."

It further observed evidence of threat actors erasing ".access" logs on "multiple" appliances in a bid to cover up the forensic trail and fly under the radar.

But by checking the artifacts that were created when triggering the SSRF vulnerability, the company said it was able to detect 670 compromised assets during an initial scan on February 3, a number that has dropped to 524 as of February 7.

In light of the continued exploitation of Ivanti devices, it's highly recommended that "all customers factory reset their appliance before applying the patch to prevent the threat actor from gaining upgrade persistence in your environment."