Mercor AI Startup Hit by Supply Chain Attack via LiteLLM Compromise
Learn More
Mercor, a San Francisco-based AI recruiting startup, reports a data breach from a supply chain attack targeting the open-source project LiteLLM.
The breach traces to late March 2026, when TeamPCP compromised the Trivy vulnerability scanner via a misconfigured GitHub Actions workflow, obtaining the PyPI publishing token for LiteLLM and uploaded two malicious versions to the public registr and adding malware designed to steal SSH keys, .env files, cloud provider credentials, cryptocurrency wallets, and AI API keys.
The Lapsus$ group has since claimed responsibility for a massive data theft from Mercor, listing 4TB of data for auction on the dark web.
The allegedly exposed data includes:
- 939GB of source code,
- 211GB user database containing resumes and personal data,
- 3TB of storage buckets (video interviews and identity verification documents),
- TailScale VPN data,
- KYC (Know Your Customer) documents.
The exact number of affected individuals has not been disclosed. Mercor claims that it contained and remediated the security incident. They are now conducting an investigation with the support of leading third-party forensic experts.
Other details are not available.