The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @moonpay/cli package globally via the npm package manager to provide the necessary tools for authentication and wallet management.\n- [COMMAND_EXECUTION]: The skill executes commands using the npm, mp (MoonPay CLI), and gog (Google Workspace CLI) tools to perform setup, verify authentication, and manage wallet lifecycle.\n- [CREDENTIALS_UNSAFE]: The documentation includes examples of importing wallets by passing mnemonic phrases and private keys directly as command-line arguments to the mp tool. This practice can result in secrets being exposed in system process lists or saved in shell history files.\n- [DATA_EXFILTRATION]: The skill references and accesses sensitive configuration directories located at ~/.config/moonpay/, which contain encrypted wallet data and user authentication tokens necessary for its operations.\n- [PROMPT_INJECTION]: The 'Autonomous login' feature introduces an indirect prompt injection surface by instructing the agent to read content from an external email account using the gog tool.\n
Ingestion points: External data enters the agent context via email search results retrieved using the gog gmail search command in SKILL.md.\n
Boundary markers: No delimiters or specific instructions are provided to separate the untrusted email content from the agent's instructions.\n
Capability inventory: The skill allows for the execution of CLI commands and access to sensitive local wallet configurations.\n
Sanitization: The skill lacks mechanisms to sanitize or validate the content of the emails before extracting the verification code.

moonpay-auth