Stolen Context.ai OAuth Token Used to Access Vercel Environments
Summary
Vercel says a stolen OAuth token linked to Context.ai gave an attacker access to some environments and variables not marked sensitive. The companies are investigating; Vercel recommends credential rotation and other safeguards.
Key points
- Threat actors reportedly used a stolen OAuth token associated with a Vercel employee’s Context.ai account, which had been granted broad “Allow All” permissions.
- Vercel says the attacker accessed some environments and variables not marked “sensitive”; it has no evidence that sensitive-marked variables were accessed.
- Vercel identified a limited subset of customers whose credentials were compromised and recommended immediate credential rotation.
- Vercel is working with Mandiant and other parties; it is investigating whether data was exfiltrated. Its services remain operational.
- Context.ai linked unauthorized access to its AWS environment to its deprecated Office Suite product, shut down that environment, and said its Bedrock platform is unaffected.
- Context.ai said the actor likely compromised OAuth tokens for some consumer users; the scope remains under investigation.
- Vercel recommends reviewing activity logs and deployments, rotating credentials and environment variables, and checking deployment-protection settings. Experts also recommend admin-managed OAuth consent and least privilege.
Article Details
- Victim Organization
- Vercel
- Incident Type
- Third-party supply-chain breach using stolen OAuth tokens to access Vercel environments.
- Data Types Exposed
- Environment variables not marked as sensitive
- Affected Records
- Unknown
- Affected Data Size
- Unknown
- Operational Impact
- Some Vercel environments and non-sensitive-marked environment variables were accessed. Vercel reported no evidence that variables marked sensitive were accessed; investigation into whether data was exfiltrated was ongoing. Vercel said its services remained operational.
- Ransom Or Extortion
- A threat actor allegedly identified as ShinyHunters was reportedly selling the breach for $2 million; the source does not confirm the sale or extortion.
- Claim Status
- confirmed
MITRE ATT&CK
People
David LindnerChief information security officer of Contrast Security; commented on the breach and the alleged sale.Guillaume ValadonCybersecurity researcher at GitGuardian; commented on identity, credentials, and SaaS integrations in attacks of this kind.Jaime BlascoChief technology officer at Nudge Security; discussed OAuth consent and the risks of stolen OAuth tokens.
Threat Actors
Vendors
Context.aiThreat actors compromised AI tool vendor Context.ai and used stolen OAuth tokens to breach software security vendor Vercel, accessing some environments and environment variables that were not marked as sensitive.CrowdStrikeWhile the company engaged CrowdStrike, conducted an investigation, closed the AWS environment, and took steps to fully deprecate the associated Office Suite product, Context learned through Vercel’s breach andMandiantThe incident underscores OAuth tokens and unsanctioned “shadow AI” tools as a growing attack surface and prompted Vercel to work with Mandiant while urging credential rotation and stronger OAuth governance.VercelThreat actors compromised AI tool vendor Context.ai and used stolen OAuth tokens to breach software security vendor Vercel, accessing some environments and environment variables that were not marked as sensitive.