The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses several shell commands to detect the environment and validate Docker configurations. These include docker info, docker ps, docker build, and find. These operations are limited to project discovery and are standard for a Docker-focused developer tool.
[DATA_EXPOSURE]: The skill demonstrates safe secrets management practices using Docker secrets and BuildKit secret mounts (e.g., /run/secrets/db_password). It does not hardcode credentials or attempt to exfiltrate sensitive files from the host system.
[EXTERNAL_DOWNLOADS]: The skill references well-known and trusted container images such as Alpine Linux, official Node.js images, Postgres, and Google's Distroless images. These references are used for instructional purposes and represent standard, safe dependencies.
[INDIRECT_PROMPT_INJECTION]: The skill processes local project files (Dockerfiles and Compose files) and Docker CLI output. While this represents an ingestion surface for untrusted data, the skill includes built-in instructions to prioritize safe internal tools and validation steps, reducing the risk of accidental execution of malicious instructions embedded in project files.

docker-expert