Developer platform WorkOS disclosed a data breach in which threat actors accessed business contact records and summaries of customer communications through an internal sales application. The attackers allegedly used an AI-powered offensive security framework to automate reconnaissance and exploitation attempts, but WorkOS said its application data and operational services were not accessed or affected.
WorkOS provides APIs, software development kits, and tools that help developers add enterprise authentication, single sign-on (SSO), user management, and identity management to their applications. According to the company’s website, at least 3,000 companies use its services, including OpenAI, Anthropic, Cursor, Perplexity, Snowflake, Plaid, Webglow, and Netlify.
WorkOS sales application compromised
“On October 30, 2026, we became aware of unauthorized access to an internal WorkOS application used by our sales team,” the company said in a written incident alert.
The information accessed by the attackers included:
- Business contact details, including names, email addresses, job titles, and companies recorded in the CRM system
- Customer account history, including product usage and connection or directory health
- Summaries and excerpts of customer sales calls, support discussions, and Slack conversations
- Business account and relationship information, including account history, sales details, and customer relationship data
After detecting the intrusion, WorkOS disabled external access to the affected applications and reviewed access controls for its internal tools. The company also hired an independent forensic and incident response firm to investigate the incident and determine its full impact.
WorkOS said it would notify affected customers directly. It also warned users to remain alert for phishing and social engineering attempts that could use the exposed information.
ARTEX AI linked to automated attack attempts
WorkOS said the attackers used an AI-powered offensive security framework to automate reconnaissance and exploitation attempts.
Company founder Michael Greenwich shared additional details, saying the attackers used the ARTEX AI penetration-testing suite. The framework has reportedly been observed in multiple incidents this year.
Chinese developers released ARTEX AI earlier this year as an open-source project for offensive security. The project was later taken private after being used in attacks targeting banks in South Korea, although several forks of its code remain available.
CrowdStrike recently reported that the cyberespionage group Vapor Panda used the penetration-testing framework in a global campaign targeting 200 organizations. The nonprofit cybersecurity organization DIVD previously said the same framework was involved in an attack reported last September.
Greenwich explained that the research team identified user agents including artex-enrich/1.0 and norma/0.4 in web requests from the threat actors.
Greenwich said the incident has been fully contained. WorkOS plans to share additional information about the attack after its investigation is complete.
Source: www.bleepingcomputer.com