agent-bricks

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (HIGH): The skill demonstrates a significant indirect prompt injection vulnerability surface through its data ingestion and configuration tools.
  • Ingestion points: Tools like create_or_update_ka ingest documents and JSON files directly from Databricks Volumes (volume_path).
  • Boundary markers: No instructions or delimiters are provided to the agent to treat content from the volumes as untrusted, increasing the risk of obedience to embedded instructions.
  • Capability inventory: The skill possesses powerful write capabilities via create_or_update_ka and create_or_update_mas, which define the core logic and routing for production AI agents.
  • Sanitization: There is no evidence of sanitization or filtering for the content retrieved from external volumes before it is applied to the agents.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 05:18 AM