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:
- Wroba Mobile Banking Trojan Spreads to the U.S. via Texts
October 30, 2020
The Wroba mobile banking trojan has made a major pivot, targeting people in the U.S. for the first time. According to researchers at Kaspersky, a wave of attacks are taking aim at U.S. Android and iPhone users in an effort that started on Thursday. The campaign uses text messages to spread, using fake notifications for “package ...
- Lazada confirms 1.1M accounts compromised in RedMart security breach
October 30, 2020
Singapore-based online grocery platform RedMart has suffered a data breach that compromised personal data of 1.1 million accounts. An individual has claimed to be in possession of the database involved in the breach, which contains various personal information such as mailing addresses, encrypted passwords, and partial credit card numbers. RedMart customers on Friday were logged out ...
- SMS Phishing Attempts Are Riding the Presidential Election Wave
October 30, 2020
SMS-based outreach has become a standard in the political playbook, with candidates and their supporters soliciting financial support, opinions, and votes through texting with increasing frequency and sophistication. In the course of protecting enterprise endpoints, Symantec, a division of Broadcom, has turned up evidence of an increasingly prevalent scam tactic in the run-up to the ...
- CISA, FBI, and CNMF Identify a New Malware Variant: ComRAT
October 29, 2020
The Cybersecurity and Infrastructure Security Agency (CISA), the Federal Bureau of Investigation (FBI), and the Department of Defense Cyber National Mission Force (CNMF) have identified a malware variant—referred to as ComRAT—used by the Russian-sponsored advanced persistent threat (APT) actor Turla. In addition, U.S. Cyber Command has released the malware sample to the malware aggregation tool and ...
- Oracle WebLogic Server RCE Flaw Under Active Attack
October 29, 2020
If an organization hasn’t updated their Oracle WebLogic servers to protect them against a recently disclosed RCE flaw, researchers have a dire warning: “Assume it has been compromised.” Oracle WebLogic Server is a popular application server used in building and deploying enterprise Java EE applications. The console component of the WebLogic Server has a flaw, CVE-2020-14882, ...
- REvil ransomware gang claims over $100 million profit in a year
October 29, 2020
REvil ransomware developers say that they made more than $100 million in one year by extorting large businesses across the world from various sectors. They are driven by profit and want to make $2 billion from their ransomware service, adopting the most lucrative trends in their pursuit of wealth. Read more… Source: Bleeping Computer