Chainstack confirms intruder entered development and production through third-party component

This is a new development in a story we reported earlier: Chainstack confirms production breach exposing customer data and password hashes.

Chainstack said on Nov. 3, 2026 that an external attacker entered its development environment and sections of its production systems through a third-party component.

Chainstack’s notice says the intruder first reached its development environment on Oct. 26. The company said it detected the activity on Oct. 31 and ended the intruder’s access right away.

Chainstack said data reached by the intruder included organization names, account holder names, member e-mail addresses and password hashes. The company added that the intruder reached systems containing customer node credentials. It stated it has no proof anyone viewed the credentials, though it treats them as exposed as a precaution.

Chainstack has not said how many customers or organizations were affected, and it has not named the third-party component used to enter its systems. It disclosed on Tuesday that a cyber attack let an unauthorized party reach some parts of its infrastructure. It said, “That does not change our responsibility.”

The Hacker News reports the breach as part of a broad AI-powered attack campaign, and word of it emerged last week. This incident fits a growing trend in which AI agents design custom attack plans and manage different stages of an intrusion.

WorkOS put out a security bulletin on Nov. 2. On Oct. 30, the company said a sales team application saw unauthorized access. The company said the accessed data included business contact details, account and relationship information, customer communications and operational metadata. The company said no evidence indicates anyone reached its production systems, and the same was true of its application data.

Last week, Chainalysis revealed it detected an outside actor. The actor employed offensive agentic techniques on its systems.

As of Nov. 2, 2026, Chainalysis had taken its Reactor and Data Solutions products offline as a precaution. The company disclosed a security incident following the use of offensive agentic techniques by attackers to reach its internal systems and tools.

Chainalysis says customer data reached by the intruders included names and email addresses. User IDs and organizational role labels were involved. The company claims no sign of exposed authentication credentials. It also claims no sign of exposed sensitive investigation data.

Chainalysis is still determining which records intruders reached. A full breakdown will follow the analysis. A spokesperson for the firm confirmed the incident and said the company is working with outside cybersecurity experts.

The firm is also using large language models (LLMs) to support a broader security review of its platform. It says it will give affected customers more updates as the investigation moves forward.

Chainalysis tells customers to enable two-factor authentication (2FA) and watch for unexpected password reset requests, login prompts and people posing as Chainalysis, particularly if they offer fund recovery or recruiting help. Chainalysis has not shared technical details of the breach, including the vulnerabilities exploited, the internal tools accessed or how long the attackers had access.

The term “agentic” generally describes AI systems that perform tasks over multiple steps with minimal human oversight, potentially allowing attackers to automate portions of an intrusion. It is still unclear whether attackers relied on existing agentic offensive security frameworks such as ARTEX or created custom AI agents that can independently find vulnerabilities, exploit them and link later actions.

Microsoft says ARTEX has recently shown up in attacks carried out by threat actors. ARTEX figured in the ConnectWise supply chain compromise.

The Hacker News says the incident overlaps with a China-linked cluster campaign targeting high-value companies using ARTEX. The Dutch Institute for Vulnerability Disclosure, known as DIVD, suffered an ARTEX breach on Sept. 21, 2026.

The Sept. 21, 2026, breach included two zero-day flaws in Zammad. The open-source ticketing system’s flaws carry the labels CVE-2026-102489 and CVE-2026-102490. The flaws let attackers raise privileges, execute code remotely and steal data.

Chainstack tells customers to change passwords, rotate node credentials and create new API keys. Customers should change their Chainstack password and change it on any other service where they reused the same password. Customers should generate new platform API keys. Those keys manage account and infrastructure.