The Agent Skills Directory

[CREDENTIALS_UNSAFE] (SAFE): The skill uses placeholder and standard development credentials (e.g., 'postgres:postgres') for template purposes. It explicitly includes warnings against hardcoding real production secrets.
[EXTERNAL_DOWNLOADS] (SAFE): Uses official and reputable Docker images (Node.js, Postgres, Redis). It follows security best practices by recommending specific version tags over the ':latest' tag.
[COMMAND_EXECUTION] (SAFE): The shell commands provided are standard administrative and debugging commands for Docker. There are no automated or suspicious script executions.
[PROMPT_INJECTION] (SAFE): The skill content is purely instructional and contains no patterns designed to manipulate the AI agent's behavior or bypass constraints.
[DATA_EXFILTRATION] (SAFE): No network exfiltration or unauthorized file access patterns were detected.

docker-patterns