The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/mp-mediaserver.py uses importlib.import_module to load media server provider classes at runtime. This is an intentional design for modularity and is safe in this context because the module names are retrieved from a hardcoded internal mapping (PROVIDER_CLASSES) and not from user-supplied input.
[DATA_EXPOSURE]: The skill interacts with local configuration databases that store media server credentials. However, the script implements a dedicated redaction layer (_is_sensitive_key and _jsonable) that identifies and strips sensitive keys (e.g., passwords, tokens, API keys) from the data before it is returned to the agent. This prevents accidental exposure of credentials.
[COMMAND_EXECUTION]: The skill requires the execute_command tool to run its bundled Python helper. The execution is scoped to a specific script provided with the skill, and the skill instructions explicitly forbid the agent from passing arbitrary URLs or credentials as arguments, adhering to the principle of least privilege.

mediaserver-operation