setting-up-a-data-warehouse-source
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements strong security practices by explicitly directing the agent to use secure, browser-based OAuth or connection links (data-warehouse-source-connect-link) for sensitive information. It specifically warns against collecting database passwords, API keys, or OAuth tokens directly within the chat interface.
- [SAFE]: The instructions include built-in security validation steps, such as checking for reachability and rejecting local or private IP addresses to prevent Server-Side Request Forgery (SSRF) attempts against the internal network.
- [SAFE]: The tools provided are focused on the stated purpose of managing data warehouse connections and do not exhibit malicious patterns like unauthorized command execution or data exfiltration.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its core functionality. Ingestion points: External database schema metadata is ingested via the external-data-sources-db-schema tool. Boundary markers: No explicit delimiters are defined, but the instructions require the agent to summarize and present the data to the user for review. Capability inventory: The skill can create data warehouse sources and register webhooks, which involves network operations. Sanitization: No programmatic sanitization is described, but the workflow requires human-in-the-loop confirmation of the selected tables and settings before any resources are created.
Audit Metadata