setting-up-a-data-warehouse-source
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes metadata from external data sources, which could be exploited by an attacker who controls those sources to influence the agent's behavior.
- Ingestion points: Data is ingested via the
external-data-sources-db-schematool, which fetches table names, column details, and descriptions from external databases (SKILL.md, Step 2). - Boundary markers: The instructions lack explicit delimiters or "ignore" directives when the agent is asked to summarize the discovered database schema for the user.
- Capability inventory: The skill allows the agent to create and configure data sources (
external-data-sources-create), manage schemas (external-data-schemas-list), and register webhooks (external-data-sources-create-webhook-create). - Sanitization: There are no instructions for sanitizing or validating the external metadata (e.g., table descriptions) before it is included in the agent's prompt context.
Audit Metadata