The Agent Skills Directory

[SAFE]: The skill provides instructional guidelines for an interactive design process. It does not contain executable code, remote script downloads, or obfuscated content.
[DATA_EXPOSURE]: While the skill instructs the agent to verify environment metadata and existing table structures, these actions are limited to gathering context for design decisions and do not involve exfiltration to external or untrusted domains.
[COMMAND_EXECUTION]: The instructions explicitly forbid the agent from generating DDL, writing scripts, or executing data pipelines directly. It restricts the agent's role to producing a design record for use by other specific tools, reducing the risk of unauthorized command execution.

data-grill