The Agent Skills Directory

[COMMAND_EXECUTION]: The main script, libtv_video.py, uses the subprocess module to manage background polling and to execute a helper script, feishu_send_video.py, for result delivery. These commands are invoked using list-based arguments rather than string shells, which is a secure practice to prevent command injection.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a relay, passing user descriptions to external image and video generation APIs.
Ingestion points: The message parameter in the create-session command within scripts/libtv_video.py.
Boundary markers: None; the skill relays the user prompt verbatim to the backend service as per its design.
Capability inventory: The skill possesses network capabilities for API interaction and result downloading, as well as local process execution for delivery automation.
Sanitization: User input is not sanitized, as the skill's primary function is to relay descriptions directly to the generation engine.

libtv-video