Chainstack says an attacker using the ARTEX AI agent framework reached its production databases through a test Grafana and an unauthenticated MCP server

Chainstack · Incident disclosure

Read the source (opens in a new tab)
What happened
Infrastructure provider Chainstack updated its November 3 incident report on November 10, 2026, saying it assesses "with high confidence" that the attacker, who first got in on October 26, used ARTEX.
Key points
  • It entered through an internet-exposed test Grafana with default admin credentials, then reached an internal MCP tool server that "required no authentication and ran with a service account bound to cluster-admin."
  • Data that may have been exposed includes organization and account holder names, members' email addresses and password hashes; node credentials were reachable, so customers are asked to rotate them.
  • Chainstack found "no evidence that a software vulnerability was exploited"; the attack chained default credentials, missing authentication, over-broad permissions and long-lived credentials.
  • It now requires authentication and least privilege for internal services with privileged operations, including AI tool servers such as MCP servers, none of which holds cluster-admin rights.
Evidence
The company's own incident report, with a timeline and indicators. The ARTEX attribution rests on naming markers, tools and behavior matching other firms' reports; Chainstack says this does not identify who was behind it.

See it in the latest record