Ghostcommit is a proof of concept that shows how AI assistants used to review software code can be tricked by hidden instructions embedded in images.
The academic ASSET Research Group showed that an attacker can place instructions inside an image file, point to it in an AGENTS.md file, and get an AI coding agent to follow those instructions during a later task.
A pull request is basically a formal “please review and add my changes” request that a developer sends before changes are added to the main version of a software project. Human reviewers and, increasingly, AI coding tools may review the changes before they’re accepted.
Read more…
Source: MalwareBytes Labs
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Atlassian update: “Take immediate action” to patch your Confluence Data Center and Server instances
November 2, 2023
Atlassian has released an advisory about a critical severity authentication vulnerability in the Confluence Server and Data Center. All versions of Confluence Data Center and Server are affected by this unexploited vulnerability. Atlassian Cloud sites are not impacted by this vulnerability, so if your Confluence site is accessed via an atlassian.net domain, it is not ...
- New Zealand: Money-motivated cyber attacks outnumber those carried out by nation-states
November 2, 2023
Major financially motivated cyber attacks in New Zealand have exceeded those launched by nation-states for the first time, and AI looms as an ever-greater weapon, a new report says. In its latest annual threat report, the National Cyber Security Centre said the potential impact was growing – though the number of major attacks dropped slightly, to ...
- Do government sanctions against ransomware groups work?
November 2, 2023
Earlier this year, the U.S. government imposed sanctions against Russian national Mikhail Matveev, an FBI most-wanted cybercriminal, who authorities accuse of being a “prolific ransomware affiliate” involved in cyberattacks in the United States and overseas. Authorities say Matveev played a major role in the development and deployment of the Hive, LockBit and Babuk ransomware variants, ...
- Ransomware gang HelloKitty exploits critical Apache ActiveMQ bug CVE-2023-46604
November 1, 2023
Beginning Friday, October 27, Rapid7 Managed Detection and Response (MDR) identified suspected exploitation of Apache ActiveMQ CVE-2023-46604 in two different customer environments. In both instances, the adversary attempted to deploy ransomware binaries on target systems in an effort to ransom the victim organizations. Based on the ransom note and available evidence, we attribute the activity to ...
- Investigation of Session Hijacking via Citrix NetScaler ADC and Gateway Vulnerability (CVE-2023-4966)
October 31, 2023
On Oct. 10, 2023, Citrix released a security bulletin for a sensitive information disclosure vulnerability (CVE-2023-4966) impacting NetScaler ADC and NetScaler Gateway appliances. Mandiant has identified zero-day exploitation of this vulnerability in the wild beginning in late August 2023 as well as n-day exploitation after Citrix’s publication. Mandiant is investigating multiple instances of successful exploitation of ...
- Over the Kazuar’s nest: Cracking down on a freshly hatched backdoor used by Pensive Ursa
October 31, 2023
While tracking the evolution of Pensive Ursa (aka Turla, Uroburos), Unit 42 researchers came across a new, upgraded variant of Kazuar. Not only is Kazuar another name for the enormous and dangerous cassowary bird, Kazuar is an advanced and stealthy .NET backdoor that Pensive Ursa usually uses as a second stage payload. Pensive Ursa is a ...