databricks-parsing
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references code examples and configuration from the official Databricks GitHub repository (github.com/databricks/bundle-examples). This is a trusted resource provided by the vendor for implementation guidance.
- [PROMPT_INJECTION]: The skill implements a workflow that ingests unstructured data from external files and processes it using AI functions, creating a surface for indirect prompt injection.
- Ingestion points: Documents are read from Databricks Unity Catalog Volumes via the read_files function in SKILL.md.
- Boundary markers: No delimiters or instructions to ignore embedded commands are present when document text is passed to ai_query.
- Capability inventory: The skill can execute LLM queries via ai_query and write data to Delta tables using Spark Structured Streaming.
- Sanitization: The skill lacks content validation or escaping mechanisms to prevent malicious instructions within documents from influencing the agent's behavior.
Audit Metadata