The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies extensively on the Databricks CLI (databricks experimental aitools tools) to interact with the workspace. It performs schema discovery, executes SQL queries, and submits DDL statements for metric view creation. These operations are consistent with the skill's stated purpose and use the official vendor CLI tool.
[INDIRECT_PROMPT_INJECTION]: The 'Metric View Advisor' workflow (documented in references/metric-view-advisor.md) creates a surface for indirect prompt injection by ingesting data from multiple external or user-provided sources.
Ingestion points: Data enters the agent context from user-provided .sql files, .csv/.yaml KPI files, and workspace assets fetched via databricks lakeview get (dashboards) and databricks api get (Genie spaces).
Boundary markers: The instructions do not specify the use of clear boundary markers or 'ignore embedded instructions' warnings for the data ingested from these sources.
Capability inventory: The skill possesses the capability to execute shell commands and submit SQL statements to the workspace (e.g., statement submit and query tools).
Sanitization: The instructions focus on extracting and merging metadata (comments, expressions, synonyms) but do not explicitly detail sanitization or validation routines for the ingested SQL or text content before it is used to generate deployment-ready definitions.

databricks-metric-views