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:
- CVE-2024-3400: Critical Command Injection Vulnerability in Palo Alto Networks Firewalls
April 17, 2024
On Friday, April 12, Palo Alto Networks published an advisory on CVE-2024-3400, a CVSS 10 zero-day vulnerability in several versions of PAN-OS, the operating system that runs on the company’s firewalls. According to the vendor advisory, if conditions for exploitability are met, the vulnerability may enable an unauthenticated attacker to execute arbitrary code with root privileges ...
- CVE-2024-20697: Windows Libarchive Remote Code Execution Vulnerability
April 17, 2024
An integer overflow vulnerability exists in the Libarchive library included in Microsoft Windows. The vulnerability is due to insufficient bounds checks on the block length of a RARVM filter used for Intel E8 preprocessing, included in the compressed data of a RAR archive. A remote attacker could exploit this vulnerability by enticing a target user into ...
- Botnets Continue Exploiting CVE-2023-1389 for Wide-Scale Spread
April 16, 2024
Last year, a command injection vulnerability, CVE-2023-1389, was disclosed and a fix developed for the web management interface of the TP-Link Archer AX21 (AX1800). FortiGuard Labs has developed an IPS signature to tackle this issue. Recently, their researchers observed multiple attacks focusing on this year-old vulnerability, spotlighting botnets like Moobot, Miori, the Golang-based agent “AGoent,” and ...
- Trust Wallet Issues Warning to Apple Users About Zero-Day Exploit in iMessage
April 16, 2024
Trust Wallet, a popular web3 wallet, has issued a warning to Apple users, urging them to disable iMessage due to “credible intel” regarding a zero-day exploit. The company shared the alert on X, stating that the exploit, which is being sold on the Dark Web, could potentially allow hackers to take control of users’ iPhones without ...
- From Social Engineering to DMARC Abuse: TA427’s Art of Information Gathering
April 16, 2024
Proofpoint researchers track numerous state-sponsored and state-aligned threat actors. TA427 (also known as Emerald Sleet, APT43, THALLIUM or Kimsuky), a Democratic People’s Republic of Korea (DPRK or North Korea) aligned group working in support of the Reconnaissance General Bureau, is particularly prolific in email phishing campaigns targeting experts for insight into US and the Republic of ...
- Cisco Duo says a third-party data breach stole MFA SMS logs
April 16, 2024
Cisco Duo has confirmed some sensitive customer data was stolen after a third-party cyber-incident. In a breach notification letter sent to affected customers, Cisco Duo said that its telephony provider, which it didn’t name, was compromised on April 1 2024. Unidentified threat actors mounted a phishing attack against the third party, through which they stole login ...