Find notable cyber news and cases, enriched with sources, timelines, and signals.

Public container registry static secret extraction analysis

Technical Analysis
First reported
Last updated
Happening score
H score 76
1 unique sources, 1 articles

Summary

Hide ▲

Public container images can leak credentials within seconds, enabling near-immediate abuse of exposed secrets across Docker Hub and other registries.

Related Happenings

Docker Hub PAT reassignment flaw (CVE-2026-105445)

Vulnerability
H score1 First: 13.10.2026 18:20 Last: 13.10.2026 18:20 Sources 1

About this happening: Docker Hub disclosed CVE-2026-105445, a critical vulnerability that let an attacker reassign a personal access token (PAT) and act as another user at the registry leve...

Developer environments using KICS data exposed after Checkmarx breach

Data Leak
H score40 First: 23.04.2026 19:05 Last: 23.04.2026 19:05 Sources 1

About this happening: The compromised Checkmarx KICS toolchain was used to exfiltrate GitHub tokens, cloud credentials, and other secrets from developer environments, creating immediate acc...

TeamPCP infostealer in compromised Trivy Docker Hub images

Malware Activity
H score37 First: 23.03.2026 17:05 Last: 23.03.2026 17:05 Sources 1

About this happening: TeamPCP infostealer was found in additional compromised Trivy Docker images, extending the malware distribution path through Docker Hub. The newly identified tags 0....

Timeline

  1. 17.11.2026 23:27 2 articles · 1d ago

    Public Docker Hub images expose AWS, OpenAI, and Anthropic keys within seconds

    Technical Analysis Update

    Stony Brook University researchers Billy Tsouvalas and Nick Nikiforakis showed that credentials embedded in public Docker Hub images can be extracted and abused within seconds after publication. In a four-month experiment spanning 300 honeypot images across eight public container registries, they planted AWS, OpenAI, and Anthropic API keys plus decoy SSH and database credentials, then observed more than 75,000 image pulls, over 140,000 AWS API calls from 5,000 unique IP addresses, and exhaustion of the test-account balances tied to the OpenAI and Anthropic keys. The researchers concluded that the scanners were overwhelmingly using static analysis of image layers or registry APIs rather than running the containers, and recommended pre-push secret scanning, publisher warnings, and rate-limiting of recently published images.

    Show sources