cloud-docker-api-abuse

Warn

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: MEDIUMPRIVILEGE_ESCALATIONPERSISTENCECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill provides explicit instructions for escaping container environments to gain root access on the host system. It details the use of bind-mounts (-v /:/host) combined with chroot /host to bypass the container boundary and operate directly on the host OS.
  • [PERSISTENCE]: The instructions include a step to establish persistent access by injecting an SSH key into the host's /root/.ssh/authorized_keys file.
  • [COMMAND_EXECUTION]: The skill outlines the execution of host-impacting commands, such as mounting host disks (mount /dev/sdaN /mnt) and manipulating Docker daemon settings via DOCKER_HOST environment variables to control containers and host resources.
  • [DATA_EXFILTRATION]: The method describes techniques for accessing sensitive host information, including reading /etc/shadow, pulling Windows NTDS/SAM files, and inspecting container image layers and environment variables for credentials.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 20, 2026, 11:02 PM
Security Audit — agent-trust-hub — cloud-docker-api-abuse