The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/validate.py correctly implements yaml.safe_load() for parsing the assets/config.yaml file, which is a security best practice to prevent arbitrary code execution during deserialization.
[SAFE]: The SQL snippets and documentation provided in SKILL.md and the references/ directory focus on legitimate data engineering practices, such as dimensional modeling and Snowflake optimization, without attempting to access sensitive system files or credentials.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided task descriptions as documented in references/GUIDE.md.
Ingestion points: User prompt input used to describe warehousing tasks.
Boundary markers: None present in the instructions.
Capability inventory: The skill primarily provides information and templates; it does not include tools or scripts with system-level write access or network capabilities that could be exploited.
Sanitization: Relies on standard LLM safety guardrails.

data-warehousing