The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides standard administrative commands for checking Docker status, system version, and managing configuration files (e.g., docker version, jq, sed). These commands are necessary for the skill's stated purpose of Docker maintenance.
[DATA_EXFILTRATION]: While the skill reads system configuration files like /etc/docker/daemon.json and /etc/os-release, it contains explicit privacy guidelines stating that it does not collect or upload user data, passwords, or credentials.
[INDIRECT_PROMPT_INJECTION]: The skill processes external information such as registry URLs and system command outputs. It mitigates potential injection risks by instructing the agent to validate JSON structures using jq and to seek user confirmation before executing significant changes.
[METADATA_POISONING]: The skill includes an external evaluation report (evaluation-report.html) that makes quality and safety claims. These claims were treated as informational and not as authoritative proof of safety during this analysis.
[PRIVILEGE_ESCALATION]: The skill acknowledges the requirement for root privileges for installation but provides warnings against closing firewalls or overwriting existing configurations without explicit user authorization.

linuxmirrors-docker