Blockchain infrastructure provider Chainstack has confirmed a leak of customers' personal data — names, email addresses, and password hashes were exposed. The incident occurred between October 26 and 31 and became one of the first documented cases of autonomous AI agents being used for a targeted attack on crypto infrastructure.

According to the disclosed details, the attacker used an AI agent system called ARTEX to penetrate the company's infrastructure. The entry point was the Grafana monitoring platform, access to which was obtained through standard administrator credentials — a classic but still critical configuration vulnerability.

Scope of consequences and recommendations

In addition to user data, Chainstack allowed the potential exposure of access keys to blockchain nodes. The company strongly recommended that customers rotate keys and change passwords. For projects using the provider's infrastructure to deploy smart contracts and manage nodes, the compromise of such keys poses a direct threat to control over assets and operational processes.

The incident demonstrates that modern attackers are moving from manual vulnerability hunting to automated multi-agent systems capable of scaling reconnaissance and exploitation many times faster. Poor administrator credential hygiene remains one of the most exploited vectors even in companies working with blockchain infrastructure.

Cryptalist analysis: The Chainstack case signals a new stage in the evolution of cyberthreats for the crypto sector — AI agents lower the entry barrier for sophisticated attacks and increase the speed of attackers' response. For institutional participants, this is an argument in favor of auditing access configurations and implementing multi-factor authentication across all infrastructure providers. The market should expect stricter security requirements from clients and regulators.