Vercel on Wednesday said it identified additional customer accounts that were compromised in a security incident that gave unauthorized access to its internal systems, but it did not disclose how many accounts were affected.
KEY FACTS
- Discovery The company expanded its review to include more compromise indicators, network requests and environment variable read events.
- Additional accounts It found a small number of customer accounts with evidence of prior compromise that predated the incident.
- Disclosure Affected parties were notified, but the exact number of impacted customers was not released.
- Attack path The breach was tied to a compromised Context.ai account used by a Vercel employee, according to a security bulletin from Vercel.
The company said the newly identified accounts were separate from the incident that exposed its internal systems and may have been compromised earlier through social engineering, malware or another method. Vercel said the accounts were reported to the affected parties after the findings were confirmed.
Vercel said the original breach began after a Vercel employee used Context.ai, which was then compromised. The attacker allegedly took over the employee’s Google Workspace account, used it to reach the employee’s Vercel account and then moved into a Vercel environment, where internal systems were explored and non-sensitive environment variables were decrypted.
Hudson Rock separately said it found signs that a Context.ai employee had been infected with Lumma Stealer in February 2026 after searching for Roblox auto-farm scripts and game exploit executors. Vercel chief executive Guillermo Rauch said threat intelligence suggested malware was distributed to computers in search of valuable tokens such as keys to Vercel accounts and other providers.
It remains unclear whether the use of the Context AI Office Suite by Vercel employees was sanctioned or amounted to shadow AI. The suite has since been deprecated by Context.ai.
WHY IT MATTERS
The incident shows how a compromise in one cloud account can spread into other services through approved integrations and stolen credentials. It also highlights the need for rapid scoping after a breach, since attackers may move through internal systems before detection.