Mercor Recruiting Compromised via Malicious LiteLLM Supply‑Chain Injection

Mercor, an AI‑driven recruiting platform, was breached when a compromised version of the open‑source LiteLLM library was introduced into its software build process. The malicious code embedded in the library activated during compilation, creating hidden back‑doors that allowed threat actors to siphon user credentials, internal configuration files, and other sensitive recruitment data.

The attack exposed data from a substantial portion of Mercor’s client base, underscoring the danger of integrating unvetted AI components into production pipelines. Defenders must enforce strict supply‑chain hygiene: verify the provenance of third‑party AI libraries, incorporate integrity checks into CI/CD workflows, and continuously monitor for anomalous behavior stemming from compromised dependencies.

Categories: Data Breaches, AI Security & Threats, Vulnerabilities & Exploits, #AI Security & Threats

Source: Read original article