Unverified50% confidenceFactExact time
According to Infosecurity Magazine, the SANDWORM_MODE worm discovered by Socket writes malicious MCP servers into the local configurations of Claude Desktop, Cursor, VS Code Continue, and Windsurf, and harvests API keys from nine LLM vendors.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/19/2026
First Seen
Valid until: 1/17/2027
Sources
Related Entities
Related Claims
UnverifiedIndependent security researcher Syed Anas Mohi Uddin discovered the same security flaw in MCP servers belonging to Google, JPMorgan Chase, and three other organizations: a tool that fetches a given URL without verifying where the address actually points.64% similarUnverifiedPast vulnerability cases: mcp-remote once passed the authorization URL provided by the server directly to the system shell (CVE-2025-6514, CVSS 9.6); the early MCP Inspector ran an unauthenticated proxy that could execute commands (CVE-2025-49596).63% similarUnverified2010年Stuxnet蠕虫病毒利用四个Windows零日漏洞，精准识别西门子S7-315和S7-417型PLC并注入恶意梯形图代码，将离心机转速在1410赫兹与2赫兹之间反复切换63% similarUnverifiedPrompt injection is a security vulnerability unique to LLM applications, where attackers can embed malicious instructions in tool results returned by an MCP server to manipulate agent behavior.61% similarUnverifiedSynapseGuard exposes a set of hand-written parameterized tool catalog (such as blast_radius, who_to_page, path_between, recent_changes) through an MCP server, rather than letting the LLM freely write Cypher queries.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1067941API
curl https://kongchang.com/api/v1/knowledge/claims/1067941MCP
get_claim(id=1067941)