The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to inspect the local Docker environment, check container status, and validate builds.
Evidence: Use of docker info, docker ps, docker images, docker build, and docker run in SKILL.md to analyze the workspace environment and validate container logic.
[INDIRECT_PROMPT_INJECTION]: The skill reads external project files such as Dockerfiles and Compose files, which could contain instructions designed to influence the agent's behavior.
Ingestion points: SKILL.md contains logic to find and read Dockerfile, *compose*.yml, and .dockerignore files from the project directory.
Boundary markers: The skill does not implement specific delimiters or warnings to ignore instructions potentially embedded within these project files.
Capability inventory: The skill has the capability to build images, run containers in the background, and execute commands within them (docker exec), providing a significant attack surface if malicious instructions are followed.
Sanitization: No sanitization, validation, or filtering of the content of the discovered files is implemented before processing.

docker-expert