Cellebrite’s forensic applications do not include the type of security protections one would expect from a parsing software, which renders them susceptible to attacks, according to privacy-focused messaging service Signal.
The Israel-based mobile forensics company offers data extraction and analysis services to intelligence organizations and public safety entities, but also to military and enterprise sectors. Cellebrite claims to have thousands of customers in over 140 countries. It has reportedly helped the FBI access information on locked phones, including in high-profile cases, but it has also been accused of providing its services to authoritarian regimes.
The company’s software solutions, Universal Forensic Extraction Device (UFED) and Physical Analyzer, work by parsing data from devices. With that data generated by the applications running on the device, Cellebrite’s software is not in control of the data, thus prone to attacks.
This “untrusted” data, Signal explains, comes in a variety of formats, depending on the applications that generated it, and could be formatted in such a way that would exploit any types of vulnerabilities in the parsing software, such as memory corruption.
“Looking at both UFED and Physical Analyzer, though, we were surprised to find that very little care seems to have been given to Cellebrite’s own software security. Industry-standard exploit mitigation defenses are missing, and many opportunities for exploitation are present,” Moxie Marlinspike, the creator of Signal, writes in a blog post.
Signal was able to execute code on a Cellebrite machine by including “a specially formatted but otherwise innocuous file” in an application running on a device that is subsequently plugged into and scanned by Cellebrite.
“There are virtually no limits on the code that can be executed,” Marlinspike notes.
One of the possible outcomes of such an attack would be to modify Cellebrite reports in unexpected manners. The attack would tamper not only with the report being generated, but also with previous and future reports, from all of the devices that were previously scanned.
Modifications may include inserting or removing contacts, emails, files, photos, texts, or any other data. According to Signal, the tampering would not result in “detectable timestamp changes or checksum failures.” Such changes could be done randomly, raising the question of the data integrity of Cellebrite’s reports.
“Any app could contain such a file, and until Cellebrite is able to accurately repair all vulnerabilities in its software with extremely high confidence, the only remedy a Cellebrite user has is to not scan devices,” Marlinspike says.
To reduce the risk, Cellebrite could update the software to not scan applications considered high risk, but even that won’t guarantee the integrity of reports.
Signal also published a proof of concept video to show what happens when Cellebrite’s UFED encounters a file designed to exploit it for arbitrary code execution.