delta-live-tables

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Data Exposure & Exfiltration (SAFE): The templates use standard cloud storage mount points (e.g., /mnt/landing/, /mnt/dlt/) for data processing. No sensitive local file paths (such as ~/.ssh or .env) are accessed, and no credentials are hardcoded. All storage paths are used as configuration parameters.
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The configuration file templates/dlt_pipeline_config.yaml references standard, versioned Python packages (pandas==1.5.3, great-expectations==0.15.50) and native Databricks init scripts. These are standard features for cluster customization and do not point to untrusted remote sources.
  • Indirect Prompt Injection (LOW): As a data processing skill, it ingests data from external storage. The skill implements robust mitigation surface via @dlt.expect decorators for data quality and schema enforcement through Spark's Auto Loader. No capabilities were identified that would allow ingested data to be executed as logic or instructions.
  • Dynamic Execution (SAFE): The skill uses a factory pattern in templates/bronze_ingestion_template.py to dynamically define DLT tables. This is a common and safe pattern in the Delta Live Tables framework for code reusability and does not involve the execution of untrusted external strings as code.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:21 PM