The Agent Skills Directory

[SAFE]: No malicious code or patterns were detected. The skill is composed of instructional markdown files that adhere to security best practices for credential management and software installation.
[COMMAND_EXECUTION]: The skill guides the agent to interact with the Databricks CLI for workspace management and data querying. It establishes clear security protocols, such as requiring explicit user selection of authentication profiles and forbidding automated profile switching.
[EXTERNAL_DOWNLOADS]: Installation instructions refer to official and trusted sources, including the Databricks GitHub repository, Homebrew, and WinGet. The guidelines emphasize manual verification of checksums and discourage piping unverified remote scripts into the shell.
[PROMPT_INJECTION]: The skill interacts with external data from Databricks tables, which creates an indirect prompt injection surface where malicious metadata could influence agent behavior.
Ingestion points: Data returned from discover-schema and query commands as described in data-exploration.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are defined for processing remote data.
Capability inventory: The agent can execute shell commands via the databricks CLI to manage resources and query data.
Sanitization: No specific data validation or sanitization steps are defined for the content retrieved from the Databricks API.

databricks-core