Security researchers have uncovered a swathe of serious vulnerabilities in a firewall system developed by CUJO which has been designed to prevent cyberattackers from infiltrating home networks.
The vulnerabilities were discovered by Claudio Bozzato from the Cisco Talos cybersecurity team.
On Tuesday, the organization published an in-depth examination of the security flaws, which included critical remote code execution bugs.
CUJO’s Smart Firewall is a device which proclaims protection from malware, viruses, and hacking attempts by way of an intelligent home firewall. However, according to Cisco Talos, the device contained flaws which were severe enough to enable cyberattackers to bypass safe browsing functions and fully hijack the product at will.
There are two main ways that hackers were able to compromise CUJO — through the execution of arbitrary code in the context of a root user, or by uploading and executing unsigned kernels on affected systems.
By combining bugs into one of two chains, attackers are also able to remotely execute code without any form of authentication.
The first chain relates to a vulnerability in the Webroot BrightCloud SDK, CVE-2018-4012, which is used by CUJO to retrieve website trust and reputation information as part of the firm’s safe browsing protection.
The bug is present in the SDK which uses HTTP — rather than an encrypted channel — to communicate with remote BrightCloud services. Should an attacker choose to intercept this traffic with a basic Man-in-The-Middle (MiTM) attack, it is what Talos describes as a “trivial” task to impersonate the BrightCloud service and execute code on the device as the root user.
CVE-2018-4031 is another security flaw impacting the safe browsing element of CUJA software. The product uses the Lunatik Lua engine to analyze network traffic, but a script injection vulnerability present in the engine allows any user, without authentication, to execute Lua scripts in the kernel.
However, as an unsafe load function is also in play, the bug can also be used to execute arbitrary code and force CUJO to extract and analyze arbitrary hostnames.
See also: Facebook debuts AI tool to tackle revenge porn
The vulnerabilities described can be chained with another security flaw, CVE-2018-3968, a boot bypass issue, to permanently hijack a CUJO device.
This security flaw is present in Das U-Boot’s Verified Boot loader. Versions 2013.07-rc1 to 2014.07-rc2 of the open-source software contain an issue which allows a boot to take place from legacy, unsigned images.
However, U-Boot is unmodifiable by design, and so CUJO cannot fix it — but the company can mitigate the problem by resolving its own bugs.
TechRepublic: How to prevent spear phishing attacks: 8 tips for your business
Two other vulnerabilities of note which have been disclosed are CVE-2018-3985 and CVE-2018-4003, both of which are code execution bugs in CUJO. While CUJO does constrain one of the processes involved which would force an attacker to escalate their privileges to utilize these flaws, if they manage to do so, this could result in arbitrary code execution on the local network.
In total, 11 security flaws were found in CUJO Smart Firewall version 7003 and reported by Cisco Talos. Static DHCP hostname command injection, record parsing errors, and denial-of-service (DoS) vulnerabilities were also disclosed.
The CUJO case serves as a reminder that if you are going to hang your hat on the promise of being a cybersecurity product, you should keep your own security practices up to par. One or two vulnerabilities may be easily forgiven — considering that every vendor in the world will eventually miss a bug or two — but this amount for a cybersecurity product seems excessive.
However, CUJO did respond to Cisco Talos’ findings quickly and has begun rolling out a system update to resolve the vulnerabilities. Users should check their devices and make sure that the latest update has been applied.
CNET: Facial recognition overkill: How deputies cracked a $12 shoplifting case
This is not the first time that CUJO has been connected to security problems. In January, failures in the CUJO API were revealed together with proof-of-concept (PoC) code which, if exploited, could result in the creation of new schedules, Internet connection tampering and DoS attacks against all devices connected to the system.
Previous and related coverage