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:
- Microsoft reveals GoldMax, Sibot and GoldFinder new malware strains used by SolarWinds hackers
March 4, 2021
Microsoft has revealed information on newly found malware the SolarWinds hackers deployed on victims’ networks as second-stage payloads. The company now tracks the “sophisticated attacker” who used the Sunburst backdoor and Teardrop malware during the SolarWinds supply-chain attack as Nobelium. Security researchers with the Microsoft Threat Intelligence Center (MSTIC) and the Microsoft 365 Defender Research Team found ...
- New Sunshuttle Second-Stage Backdoor Uncovered Targeting U.S.-Based Entity; Possible Connection to UNC2452
March 4, 2021
Mandiant Threat Intelligence discovered a new backdoor uploaded by a U.S.-based entity to a public malware repository in August 2020 that we have named SUNSHUTTLE. SUNSHUTTLE is written in GO, and reads an embedded or local configuration file, communicates with a hard-coded command and control (C2) server over HTTPS, and supports commands including remotely uploading ...
- Threat Assessment: Active Exploitation of Four Zero-Day Vulnerabilities in Microsoft Exchange Server
March 4, 2021
On Mar. 2, 2021, Volexity reported in-the-wild-exploitation of four Microsoft Exchange Server vulnerabilities: CVE-2021-26855, CVE-2021-26857, CVE-2021-26858 and CVE-2021-27065. As a result of these vulnerabilities being exploited, adversaries can access Microsoft Exchange Servers and allow installation of additional tools to facilitate long-term access into victims’ environments. There has also been a report of multiple threat actors leveraging ...
- Supermicro, Pulse Secure release fixes for ‘TrickBoot’ attacks
March 4, 2021
Supermicro and Pulse Secure have released advisories warning that some of their motherboards are vulnerable to the TrickBot malware’s UEFI firmware-infecting module, known as TrickBoot. Last year, cybersecurity firms Advanced Intelligence and Eclypsium released a joint report about a new malicious firmware-targeting ‘TrickBoot’ module delivered by the notorious TrickBot malware. When executed, the module will analyze a ...
- Maza Russian cybercriminal forum suffers data breach
March 4, 2021
The Maza cybercriminal forum has reportedly suffered a data breach leading to the leak of user information. On March 3, Flashpoint researchers detected the breach on Maza — once known as Mazafaka — which has been online since at least 2003. Maza is a closed and heavily-restricted forum for Russian-speaking threat actors. The community has been connected ...
- Google patches actively exploited Chrome browser zero-day vulnerability
March 3, 2021
Google has warned of reports that a zero-day vulnerability in the Chrome browser is being actively exploited in the wild. The vulnerability, tracked as CVE-2021-21166, was reported by Alison Huffman from the Microsoft Browser Vulnerability Research team on February 11 and is described as an “object lifecycle issue in audio.” Google has labeled the vulnerability as a ...