Gumroad Automation
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Gumroad API resources, including product names, customer emails, and license verification statuses.
- Ingestion points: Data retrieved via
GUMROAD_LIST_PRODUCTS,GUMROAD_GET_SALES, andGUMROAD_VERIFY_LICENSE(SKILL.md). - Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions to separate processed data from agent instructions.
- Capability inventory: The skill is scoped to Gumroad API operations; it does not have access to arbitrary shell execution, file system modifications, or general network requests beyond the defined MCP integration.
- Sanitization: No explicit validation or escaping of strings retrieved from the Gumroad API is documented.
Audit Metadata