The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the mp (MoonPay) CLI tool to perform paid API requests. This involves shell command execution where the agent must interpolate external URLs and JSON bodies as arguments.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data that influences high-stakes financial operations, creating an indirect prompt injection vulnerability surface.
Ingestion points: The --url and --body parameters in the mp x402 request command defined in SKILL.md.
Boundary markers: There are no protective boundary markers or instructions to the agent to validate or ignore malicious commands embedded within the fetched content or user-supplied bodies.
Capability inventory: The CLI tool mp is capable of signing and broadcasting blockchain transactions, allowing the agent to spend funds from a local wallet.
Sanitization: No sanitization or schema validation steps are defined for the JSON body or the target URL before they are passed to the shell command.
[DATA_EXFILTRATION]: The skill identifies and transmits wallet names or addresses to external endpoints (agents.moonpay.com) during the payment flow. While this is directed at vendor-managed infrastructure, it involves the transmission of sensitive financial identifiers to a remote server.

moonpay-x402