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:
- Giant Tiger breach sees 2.8 million records leaked
April 16, 2024
When asked, they posted a small snippet as proof. The download of the full database is practically free for other active members of that forum. In March, one of Giant Tiger‘s vendors, a company used to manage customer communications and engagement, suffered a cyberattack, which impacted Giant Tiger, as reported by CBC. The retailer first learned ...
- South Africa’s trade regulator ITAC hit by cyber attack
April 16, 2024
According to the organisation, the attack happened in January, leading to the exposure of the personal information of stakeholders. ITAC is an institution dedicated to promoting fair trade in South Africa in order to enhance economic growth and development. The site includes trade and tariff services as well as import and export control services. In a ...
- MGM files suit against FTC to block cyber attack investigation
April 16, 2024
MGM filed the suit yesterday (15 April) in Washington’s federal court against both the FTC and Lina M Khan as FTC chair. The suit refers to the large-scale cyber attack launched against MGM in September last year. MGM was forced to shut down certain systems across its US properties due to the attack. Access to MGM ...
- Best Practices for Deploying Secure and Resilient AI Systems
April 15, 2024
Deploying artificial intelligence (AI) systems securely requires careful setup and configuration that depends on the complexity of the AI system, the resources required (e.g., funding, technical expertise), and the infrastructure used (i.e., on premises, cloud, or hybrid). This report expands upon the ‘secure deployment’ and ‘secure operation and maintenance’ sections of the Guidelines for secure AI ...
- SteganoAmor campaign: TA558 mass-attacking companies and public institutions all around the world
April 15, 2024
Researchers from the Positive Technologies Expert Security Center discovered more than three hundred attacks worldwide, which they confidently attributed to the well-known TA558 group. As originally described by researchers at ProofPoint, TA558 is a relatively small financially motivated cybercrime group that has attacked hospitality and tourism organizations mainly in Latin America, but has also been identified ...
- Change Healthcare faces another ransomware threat – and it looks credible
April 12, 2024
For months, Change Healthcare has faced an immensely messy ransomware debacle that has left hundreds of pharmacies and medical practices across the United States unable to process claims. Now, thanks to an apparent dispute within the ransomware criminal ecosystem, it may have just become far messier still. In March, the ransomware group AlphV, which had claimed ...