Unverified50% confidenceFactExact time
The malicious extension was not developed by Anthropic but was distributed through an official account, constituting a supply chain compromise; the payload attempted to implant a persistent SSH backdoor, steal information such as hostname, username, architecture and public IP, and write it into authorized keys, failing on most machines because it lacked permission to start the service.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/1/2026
First Seen
Valid until: 11/15/2026
Sources
OpenAI, Anthropic, and Google Reveal New Moves on the Same Day: Model Iteration and Ecosystem Shifts
bilibili菲猫Feimao10/31/2026
Related Entities
Related Claims
UnverifiedWithout rigorous authentication, permission isolation, and input validation, a hijacked agent may cause damage to backend systems via an MCP server.73% similarUnverifiedIndependent 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.72% similarUnverifiedOnce secrets and tokens are scanned and exfiltrated, they may lead to hijacking of cloud service accounts, tampering with code repositories, and use as a springboard for supply chain attacks.70% similarUnverifiedThe root cause of the vulnerability was that the internal service did not properly validate the token signature and simultaneously used upn as a local lookup key.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1124479API
curl https://kongchang.com/api/v1/knowledge/claims/1124479MCP
get_claim(id=1124479)