The Agent Skills Directory

[DATA_EXFILTRATION]: The skill facilitates the handling of sensitive payment credentials. It provides functionality to write unmasked credit card details, including the full Primary Account Number (PAN), CVC, and expiration dates, to a local file (e.g., /tmp/link-card.json) using the --output-file flag. It also includes code examples for reading these credentials from the filesystem to automate checkouts.
[CREDENTIALS_UNSAFE]: Authentication with the Link wallet is managed through local credential files. The skill supports specifying these files via the --auth flag or the LINK_AUTH_FILE environment variable, which stores sensitive session tokens on disk.
[COMMAND_EXECUTION]: The skill relies on executing a series of shell commands for its core functionality, including authentication, spend request creation, polling for user approval, and executing payments through the Machine Payment Protocol (MPP).
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill processes potentially untrusted data such as merchant names, URLs, and purchase context descriptions (minimum 100 characters) which are passed as arguments to the CLI.
Boundary markers: None identified; input data is directly interpolated into command-line arguments.
Capability inventory: The skill has the ability to execute shell commands, write sensitive financial data to the filesystem, and perform network requests via the MPP pay command.
Sanitization: The instructions do not specify sanitization or escaping for the input strings provided to the spend request commands.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @stripe/link-cli package from the npm registry. This package is hosted within a well-known organization's official namespace.

stripe-link-cli