The Agent Skills Directory

[SAFE]: The skill implements secure credential management by using getpass to prevent password echoing and by applying restricted file permissions (chmod 600) to its local configuration file (~/.pandadata/pandadata.env), ensuring that sensitive data is only readable by the owner.
[COMMAND_EXECUTION]: Subprocess calls are used for environment management, such as installing dependencies via pip and launching interactive setup terminals. These operations are scoped to the skill's internal functionality and use safe path construction methods.
[EXTERNAL_DOWNLOADS]: The skill installs the panda-data package and its dependencies from the official PyPI registry. This is required for its primary purpose and uses standard package management practices.
[PROMPT_INJECTION]: The instructions in SKILL.md prioritize the provided local documentation over the model's training data and explicitly forbid the fabrication of API parameters or authentication steps, which helps ensure the reliability and safety of generated code.
[DATA_EXFILTRATION]: Network activity is restricted to making API requests to a Pandadata service URL provided by the user, which represents the intended functionality for a financial market data interface.

pandadata-api