The Agent Skills Directory

[COMMAND_EXECUTION]: Executes a standard docker run command to launch the Metabase semantic checker. The command includes mounting the current working directory to the container to allow the tool to access local configuration files.
[EXTERNAL_DOWNLOADS]: Pulls the official metabase/metabase-enterprise Docker image from Docker Hub to perform the validation.
[PROMPT_INJECTION]: The skill processes user-controlled YAML files and database metadata JSON. This constitutes an indirect prompt injection surface where maliciously crafted data could attempt to influence the agent's behavior, although the risk is mitigated by the tool running within an isolated Docker container.

metabase-semantic-checker