The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands for core operations such as project initialization, deployment, and testing.
Uses databricks bundle init with process substitution to scaffold projects.
Recommends uv run pytest for local execution of unit tests.
Instructs the agent to manage workflows using CLI commands like databricks bundle deploy and databricks jobs run-now.
[EXTERNAL_DOWNLOADS]: The skill guides the user to install the Databricks CLI from external sources with appropriate security warnings.
References official Homebrew taps and GitHub releases for tool installation.
Explicitly warns against piping remote installers directly into a shell and recommends verifying binary checksums before installation.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external data and parameters.
Ingestion points: Data enters the agent's context through dbutils.widgets.get() in notebooks and dbutils.jobs.taskValues.get() in job triggers, which can pull from SQL tables or file arrival metadata.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the provided templates to isolate untrusted data.
Capability inventory: The skill can trigger job runs, deploy infrastructure bundles, and execute SQL queries within a Databricks environment.
Sanitization: Reference examples in the documentation demonstrate SQL query construction using string interpolation (f-strings), which lack sanitization against potentially malicious input stored in parameters.

databricks-jobs