The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Stripe (e.g., customer names, metadata, and documentation) which could potentially contain malicious instructions intended to influence the agent.
Ingestion points: Tools such as list_customers, search_stripe_resources, and search_stripe_documentation ingest external content from Stripe's systems (tools.md).
Boundary markers: The skill does not define explicit delimiters or instructions to ignore commands embedded within the data retrieved from Stripe (SKILL.md).
Capability inventory: The skill has significant capabilities, including create_refund, cancel_subscription, and create_invoice, which could be targeted by an injection attack (SKILL.md, tools.md).
Sanitization: There is no mention of sanitizing or validating the content fetched from Stripe before it is used in the agent context.
[COMMAND_EXECUTION]: The skill provides instructions for executing shell commands to configure the MCP connection using the claude CLI. These commands modify the local environment and interact with external transport URLs (SKILL.md).

mcp-stripe