agent-checkouts
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface related to the ingestion of data from third-party merchant pages via the Crossmint API.
- Ingestion points: The agent is instructed to process
messageandresponseSchemafields from thependingUserActionobject inSKILL.md. - Boundary markers: No explicit delimiters or instructions are provided to the agent to treat API-returned strings as untrusted or to ignore embedded instructions.
- Capability inventory: The skill uses
scripts/checkout.mjsto make network requests to the Crossmint API, which can execute payments and order placement. - Sanitization: The skill lacks specific sanitization or filtering of text and schema definitions returned by the merchant-controlled API responses.
Audit Metadata