Skills
skills/jeremylongshore/claude-code-plugins-plus-skills/databricks-core-workflow-a/Gen Agent Trust Hub

databricks-core-workflow-a

Warn

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The function aggregate_to_gold in src/pipelines/gold.py uses the Python eval() function to process aggregation expressions constructed from string arguments. This pattern is highly dangerous as it allows for arbitrary code execution if the input dictionary keys or values are manipulated or contain malicious payloads.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates reading and transforming data from external storage paths using powerful tools, which presents a vulnerability to indirect injection if the processed data contains malicious instructions.\n
  • Ingestion points: Data ingestion occurs through source_path in ingest_to_bronze and table references in transform_to_silver and aggregate_to_gold.\n
  • Boundary markers: There are no boundary markers or explicit instructions provided to the agent to disregard embedded commands in the data being processed.\n
  • Capability inventory: The skill utilizes Bash(databricks:*), Write, Edit, and Read capabilities.\n
  • Sanitization: No data validation or sanitization routines are implemented to clean or verify external content before it is passed to downstream processing or aggregation functions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 12, 2026, 12:03 AM