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:
- ‘Fully undetectable’ Windows backdoor gets detected
October 18, 2022
SafeBreach Labs says it has detected a novel fully undetectable (FUD) PowerShell backdoor, which calls into question the accuracy of threat naming. More significantly, the malware may backdoor your Windows system by masquerading as part of the update process. Tomer Bar, director of security research at SafeBreach, explains in an advisory that the software nasty and associated ...
- Spyder Loader: Malware Seen in Recent Campaign Targeting Organizations in Hong Kong
October 18, 2022
Symantec has observed a likely continuation of the Operation CuckooBees activity, this time targeting organizations in Hong Kong. Operation CuckooBees was first documented in May 2022 by researchers at Cybereason, who said the intelligence-gathering campaign had been operating under the radar since at least 2019, stealing intellectual property and other sensitive data from victims. The victims observed ...
- CISA Releases Two Industrial Control Systems Advisories
October 18, 2022
CISA released two Industrial Control Systems (ICS) advisories on October 18, 2022. These advisories provide timely information about current security issues, vulnerabilities, and exploits surrounding ICS. CISA encourages users and administrators to review the newly released ICS advisories for technical details and mitigations: ICSA-22-291-01 Advantech R-SeeNet ICSA-21-336-06 Hitachi Energy APM Edge (Update A) Read more… Source: U.S. Cybersecurity and Infrastructure ...
- Verizon notifies prepaid customers their accounts were breached
October 18, 2022
Verizon warned an undisclosed number of prepaid customers that attackers gained access to Verizon accounts and used exposed credit card info in SIM swapping attacks. “We determined that between October 6 and October 10, 2022, a third party actor accessed the last four digits of the credit card used to make automatic payments on your account,” ...
- Defenders beware: A case for post-ransomware investigations
October 18, 2022
Ransomware is one of the most pervasive threats that Microsoft Detection and Response Team (DART) responds to today. The groups behind these attacks continue to add sophistication to their tactics, techniques, and procedures (TTPs) as most network security postures increase. In this blog, DART researchers detail a recent ransomware incident in which the attacker used a ...
- The benefits of taking an intent-based approach to detecting Business Email Compromise
October 18, 2022
Business email compromise (BEC) is one of the most financially damaging online crimes. As per the internet crime 221 report, the total loss in 2021 due to BEC is around 2.4 billion dollars. Since 2013, BEC has resulted in a 43 billion dollars loss. The report defines BEC as a scam targeting businesses (not individuals) ...