control-tower-design
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the instructions or the provided code.
- [DATA_EXPOSURE]: The example Python code uses generic placeholders for connection parameters (e.g.,
api_key: 'xxx'), which is a standard and safe way to demonstrate configuration without exposing actual credentials. - [INDIRECT_PROMPT_INJECTION]: The skill defines functions to process order and shipment data; however, these are provided as static code templates for the user. The skill does not automate the ingestion of untrusted external data into the agent's prompt context, nor does it possess dangerous tools or capabilities that could be exploited via data injection.
Audit Metadata