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:
- Attackers scan 1.6 million WordPress sites for vulnerable plugin
July 15, 2022
Security researchers have detected a massive campaign that scanned close to 1.6 million WordPress sites for the presence of a vulnerable plugin that allows uploading files without authentication. The attackers are targeting the Kaswara Modern WPBakery Page Builder, which has been abandoned by its author before receiving a patch for a critical severity flaw tracked as ...
- The industrial internet of things is still a big mess when it comes to security
July 14, 2022
Critical infrastructure is increasingly targeted by cyber criminals – and while those responsible for running industrial networks know that securing operational technology (OT) and the Industrial Internet of Things (IIoT) is vital, they’re struggling, resulting in networks being left vulnerable to attacks. According to analysis by cybersecurity company Barracuda, 94% of industrial organisations have experienced a ...
- Cyber Safety Review Board Releases Unprecedented Report of its Review into Log4j Vulnerabilities and Response
July 14, 2022
WASHINGTON – Today, the U.S. Department of Homeland Security (DHS) released the Cyber Safety Review Board’s (CSRB) first report, which includes 19 actionable recommendations for government and industry. The recommendations from the CSRB – an unprecedented public-private initiative that brings together government and industry leaders to review and assess significant cybersecurity events to better protect ...
- New Lilith ransomware emerges with extortion site, lists first victim
July 13, 2022
A new ransomware operation has been launched under the name ‘Lilith,’ and it has already posted its first victim on a data leak site created to support double-extortion attacks. Lilith is a C/C++ console-based ransomware discovered by JAMESWT and designed for 64-bit versions of Windows. Like most ransomware operations launching today, Lilith performs double-extortions attacks, which ...
- Uncovering a macOS App Sandbox escape vulnerability: A deep dive into CVE-2022-26706
July 13, 2022
Microsoft uncovered a vulnerability in macOS that could allow specially crafted codes to escape the App Sandbox and run unrestricted on the system. We shared these findings with Apple through Coordinated Vulnerability Disclosure (CVD) via Microsoft Security Vulnerability Research (MSVR) in October 2021. A fix for this vulnerability, now identified as CVE-2022-26706, was included in ...
- Transparent Tribe begins targeting education sector in latest campaign
July 13, 2022
Cisco Talos recently discovered an ongoing campaign conducted by the Transparent Tribe APT group against students at various educational institutions in India. This campaign was partially covered by another security firm, but our findings reveal more details regarding the adversary’s operations. Typically, this APT group focuses on targeting government (government employees, military personnel) and pseudo-government entities ...