Fireworks AI Hack Hits Cursor Extensions and 150K Uber Drivers
Cursor traced its tampered Claude Code extension incident to the Fireworks AI breach, while Uber is notifying about 150,000 drivers whose data was stolen via the same provider.

⚡ Why it matters
A single access key held by a third-party AI provider gave the attacker a path into other companies and, ultimately, onto developer machines.
Cursor, the company behind the AI-assisted code editor, says the late-October tampering of its extension marketplace began with a breach at Fireworks AI, a cloud platform for running and fine-tuning open source AI models. Uber and Notion have also disclosed data exposure tied to the same Fireworks compromise.
How the attacker reached Cursor
According to Cursor, as reported by SecurityWeek, Fireworks was breached twice in October. In one of those incidents the attacker obtained an access key that Cursor had issued to Fireworks. With it, the attacker copied data out of a read-only storage bucket that held unredacted credentials for SaaS tools Cursor uses.
Those tools gave access to archived non-production code repositories, a limited set of support tickets and emails, product usage metrics, summaries of some coding agent conversations, and credentials that some users had sent to Cursor support.
Two tampered Claude Code builds
The intrusion was spotted on October 29. Using secrets taken from Cursor's internal systems, the attacker placed modified copies of the anthropic.claude-code extension, versions 2.1.312 and 2.1.313, in the cache that serves Cursor's VS Code extension marketplace. They stayed available for roughly 11 hours.
- Anthropic was not breached, and the Open VSX source never contained the malicious builds.
- Cursor estimates that no more than 5,000 users were affected.
- The company has seen no attacker activity since November 1 and considers the incident contained.
Cursor warns that simply uninstalling the extension may not be enough and points affected users to its security notice.
Uber: about 150,000 drivers
Uber is separately notifying roughly 150,000 drivers. A spokesperson confirmed to TechCrunch that the third-party provider involved was Fireworks AI. Uber says:
- On October 7, attackers used the provider's compromised credentials to reach data Uber had shared with it.
- Exposed data includes driver's license images, names, addresses, dates of birth, license numbers and government ID numbers.
- Existing alerts did not fire because the attacker used valid credentials.
- A filing with the Texas Attorney General's Office puts the total at 149,346 people. Uber is offering free credit monitoring.
Fireworks first said it saw no sign that customer personal data was accessed, but a November 9 update acknowledged that account and billing details and some customer datasets were affected.
Recommendations
- If you installed or updated Claude Code to 2.1.312 or 2.1.313 from the Cursor marketplace during the exposure window, follow Cursor's security notice and rotate secrets and API keys stored on that machine.
- Review access keys granted to third-party providers, scope them tightly and rotate them regularly.
- Never send passwords or keys to a vendor's support team.
- Affected Uber drivers should enroll in the free credit monitoring and watch for related phishing.
We covered the malicious extension earlier in our first report on the Cursor incident.
🔗 Sources
This article is original reporting by the Security World editorial team based on the sources listed above; direct quotes are kept short and marked with quotation marks. Corrections and feedback: info [at] securitywd.net.
Related stories
FBI Probes Apparent Breach of NCIC National Crime Information Database
The FBI is investigating an apparent hack of NCIC, the criminal justice database used by US law enforcement agencies. The scope of the incident is still unclear.
Rackspace Spot Breach: Leaked Credentials Exposed Kubernetes Secrets
Attackers used leaked credentials to access Rackspace Spot on Nov 17, creating pods and reading Kubernetes Secrets in some customer clusters. Rotate credentials now.
Microsoft Entra Connect Health API Flaw Exposed Admin Emails and Tenant Data
Microsoft says attackers abused an Entra Connect Health API bug to read admin emails and tenant metadata. A server-side fix is live; phishing risk remains.


