The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends using official Stripe developer tools, including the Stripe CLI (@stripe/cli) and the official MCP server (mcp.stripe.com). These are well-known, trusted resources from a recognized vendor.
[COMMAND_EXECUTION]: Instructions include standard command-line operations for setting up and testing integrations, such as stripe login, stripe listen, and stripe trigger. These are appropriate and safe for the intended developer workflow.
[PROMPT_INJECTION]: The skill explicitly manages risks related to processing external data from APIs and webhooks by incorporating robust security controls.
Ingestion points: External data enters through Stripe API responses (via tools) and incoming webhook HTTP requests.
Boundary markers: Provides clear instructions to the agent to treat tool outputs as data rather than instructions to prevent autonomous command execution.
Capability inventory: The skill allows for Stripe SDK mutations (e.g., subscription creation) and local database updates based on the processed data.
Sanitization: Mandates the use of raw body cryptographic signature verification (stripe.webhooks.constructEvent) to ensure the authenticity and integrity of all external data.

stripe-billing