The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow for processing data from external databases via execute_query and get_context, which represents a theoretical surface for indirect prompt injection.
Ingestion points: The skill ingests untrusted data distributions and schema metadata through execute_query and get_context calls described in SKILL.md.
Boundary markers: Absent. The instructions do not define specific delimiters or instructions to ignore embedded commands within the database data.
Capability inventory: The skill utilizes execute_query for database interaction and writes results to modeling-notes.md.
Sanitization: Absent. There is no explicit requirement for the agent to sanitize or escape data retrieved from the database before presenting it or recording it in notes.
[COMMAND_EXECUTION]: The skill uses the execute_query tool to perform row counts and data quality checks. This behavior is consistent with the skill's primary stated purpose of facilitating data modeling and analytical scope discovery.

malloy-scope