The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as product names and descriptions provided by users or administrators. If this data is later presented to an AI agent without sanitization, it could serve as a surface for indirect prompt injection.
- Ingestion points: The createCheckoutSession function in src/backend/main.mo accepts items containing productName and productDescription.
- Boundary markers: No explicit boundary markers or instructions to ignore embedded commands were found in the data processing logic.
- Capability inventory: The skill has the ability to perform HTTP outcalls to Stripe via the OutCall module and execute state changes in the backend actor.
- Sanitization: There is no evidence of string validation or sanitization for product metadata fields before they are processed by the backend or stored.

extension-stripe