Research

Cloud keys leaked in Docker Hub images are stolen within seconds

Page published

Earliest dated coverage: 17 Nov 2026 · First observed: 17 Nov 2026 · Latest dated coverage: 17 Nov 2026

Coverage timeline

17 Nov 2026bleepingcomputer.com

Single-source research — one report is available.

Why it matters

The observed abuse of OpenAI and Anthropic keys shows that secrets exposed in container images can enable unauthorized AI API consumption and depleted account balances.

Stony Brook University researchers published 300 honeypot images across eight public container registries to measure abuse of exposed credentials, including AWS, OpenAI, and Anthropic API keys. The fastest observed AWS-key use occurred 7.86 seconds after publication, while exposed OpenAI and Anthropic keys were hijacked to exhaust test-account balances. Monitoring indicated that scanners generally extracted secrets directly from image layers without running the containers.