The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from the database and the environment to determine data freshness.
Ingestion points: Results from SQL queries (like column names from INFORMATION_SCHEMA.COLUMNS) and output from shell commands (af dags list).
Capability inventory: The skill has the ability to execute SQL queries and shell-based CLI commands (af).
Boundary markers: No explicit delimiters are used to separate data from instructions when reporting or using output to find DAGs.
Sanitization: There are no instructions for sanitizing or validating table names or DAG IDs obtained from the system before they are used in subsequent commands like af dags get <dag_id>.
[COMMAND_EXECUTION]: The skill instructs the agent to execute specific shell commands using the 'af' tool.
Evidence: Found in SKILL.md referencing af dags list, af dags get <dag_id>, and af dags stats.
Context: These commands are used to verify the health of data pipelines managed by Airflow/Astro, which is consistent with the skill's primary purpose.

checking-freshness