Verified65% confidenceFactExact time
MCP server spoofing attacks exploit the assumption in the protocol design that servers are trusted by default, rather than exploiting vulnerabilities in the traditional sense.
3
Sources
65%
Confidence
Long-term
Relevance
10/8/2026
First Seen
Sources
Related Entities
Related Claims
Unverified安全研究人员发现一个恶意MCP服务器伪装成正规邮件转发服务Postmark MCP，实际会在每封邮件密送一份给攻击者77% similarUnverifiedThe caller of an MCP server is a model that may be misled, injected or simply make mistakes, so the security design must achieve "even if it errs, the fragmenting consequences are small and visible."76% similarVerifiedSupply-chain-level MCP attack patterns include typosquatted servers and rug pulls (i.e., tool definitions that are silently altered after being approved).76% similarUnverifiedTrustFloor is implemented as an MCP gateway: after the agent calls the gateway, it executes a policy check and returns one of three results: allow, pending approval, or reject.76% similarVerifiedUnlike AgentReach, an MCP server exposes a unified protocol endpoint as an intermediate layer, meaning every call must pass through that intermediate layer, which can become a performance bottleneck or single point of failure.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/987546API
curl https://kongchang.com/api/v1/knowledge/claims/987546MCP
get_claim(id=987546)