The TeamPCP claim around Mistral AI repositories shows how software supply-chain attacks are now moving beyond package poisoning into source-code theft, extortion, and exposure of internal development workflows. Mistral AI confirmed that a codebase management system was compromised following the Mini Shai-Hulud supply-chain attack, though the company stated that hosted services, managed user data, and research/testing environments were not compromised.

For AI companies, this is especially sensitive because repositories may contain training, fine-tuning, benchmarking, inference, delivery, CI/CD, and experimental project logic. Even if core production systems are not impacted, stolen source code can reveal architecture, secrets, future product direction, security assumptions, and integration points. Organizations should urgently review developer endpoint security, CI/CD credential exposure, package publishing workflows, repository access controls, secret scanning, and token rotation. Apparently, attackers have discovered that AI companies are also just software companies with credentials, pipelines, and humans, which is devastatingly unfair to everyone’s marketing department.

The TeamPCP hacker group is threatening to leak source code from the Mistral AI project unless a buyer is found for the data. [...]

Source: TeamPCP hackers advertise Mistral AI code repos for sale via Bleeping Computer — published 14 May 2026.