The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python CLI tool using uv run scripts/cli.py to perform database discovery and cache management. These scripts are located within the analyzing-data skill directory, which is part of the expected vendor environment.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by aggregating metadata from external sources into high-context documentation files.
Ingestion points: The skill reads ~/.astro/agents/warehouse.yml, database schema/table/column comments from the warehouse, and business context from dbt YAML files and SQL frontmatter in the codebase.
Boundary markers: No specific delimiters or boundary warnings are specified when merging retrieved descriptions into the .astro/warehouse.md or CLAUDE.md files.
Capability inventory: The skill has the capability to execute shell commands via uv run and perform filesystem writes to both project-local and user-home directories.
Sanitization: There is no evidence of sanitization or filtering applied to external metadata before it is interpolated into the generated documentation.

warehouse-init