The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the configuration of an external Model Context Protocol (MCP) server at https://rube.app/mcp. This third-party service acts as the execution environment for all Stripe operations.
The service is described as requiring no local API keys, implying that authentication and credential management are handled by the remote provider.
This creates a dependency on an external infrastructure for sensitive financial operations.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Stripe (such as customer names, emails, and metadata) which could contain malicious instructions.
Ingestion points: Data is ingested through tools like STRIPE_LIST_CUSTOMERS, STRIPE_SEARCH_CUSTOMERS, and STRIPE_RETRIEVE_SUBSCRIPTION (SKILL.md).
Boundary markers: No delimiters or instructions to ignore embedded content are present in the workflows.
Capability inventory: The skill has high-impact capabilities including issuing refunds (STRIPE_CREATE_REFUND), creating charges (STRIPE_CREATE_PAYMENT_INTENT), and modifying customer records.
Sanitization: There is no evidence of sanitization or validation of the data retrieved from Stripe before it is processed by the agent.

stripe-automation