This article examines the security implications of the Model Context Protocol (MCP) sampling feature in the context of a widely used coding copilot application.
MCP is a standard for connecting large language model (LLM) applications to external data sources and tools. We show that, without proper safeguards, malicious MCP servers can exploit the sampling feature for a range of attacks. We demonstrate these risks in practice through three proof-of-concept (PoC) examples conducted within the coding copilot, and discuss strategies for effective prevention.
Read more…
Source: Palo Alto Unit 42
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Telegram has become a window into war
July 7, 2023
Since the invasion of Ukraine in February 2022, Telegram has gained an outsize influence on one of the world’s most watched conflicts. “Telegram is fantastic for many, many reasons and for the fact that we’ve managed to see what is happening at such a crucial point in history,” says Jordan Wildon, digital investigator and founder ...
- The five-day job: A BlackByte ransomware intrusion case study
July 6, 2023
As ransomware attacks continue to grow in number and sophistication, threat actors can quickly impact business operations if organizations are not well prepared. In a recent investigation by Microsoft Incident Response (previously known as Microsoft Detection and Response Team – DART) of an intrusion, they found that the threat actor progressed through the full attack ...
- Charges filed in cyber attack on East Bay water treatment plant
July 6, 2023
A 53-year-old Tracy man is facing federal criminal charges in connection with an alleged attack on the computer systems of a Discovery Bay water treatment plant more than two years ago, according to the U.S. Attorney’s Office. Rambler Gallo was a full-time employee of a private Massachusetts-based company that contracted with Discovery Bay to operate the ...
- CISA and Partners Release Joint Cybersecurity Advisory on Newly Identified Truebot Malware Variants
July 6, 2023
Today, the Cybersecurity and Infrastructure Security Agency (CISA), Federal Bureau of Investigations (FBI), the Multi-State Information Sharing and Analysis Center (MS-ISAC), and the Canadian Centre for Cyber Security (CCCS) released a joint Cybersecurity Advisory (CSA), Increased Truebot Activity Infects U.S. and Canada Based Networks, to help organizations detect and protect against newly identified Truebot malware ...
- Japan’s biggest port hit by suspected cyberattack, operations halted
July 5, 2023
The Port of Nagoya, Japan’s largest port by total cargo throughput and responsible for handling some of Toyota Motor Corp.’s car exports, has suffered a crippling system glitch, with the port operator saying Wednesday it suspects a cyberattack. As of noon, the port in central Japan remained unable to load and unload containers from trailers. Police ...
- ChatGPT Shared Links and Information Protection: Risks and Measures Organizations Must Understand
July 5, 2023
Since its initial release in late 2022, the AI-powered text generation tool known as ChatGPT has been experiencing rapid adoption rates from both organizations and individual users. However, its latest feature, known as Shared Links, comes with the potential risk of unintentional disclosure of confidential information. In this article, Trend Micro researchers will examine these risks ...