The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data via prompts and interpolates them into command-line executions.
Ingestion points: User-provided prompts passed to the --prompt argument in comfy-videogen commands as shown in SKILL.md.
Boundary markers: Absent; there are no instructions to wrap or delimit user input to prevent it from being interpreted as part of the command structure.
Capability inventory: The skill relies on uv run to execute shell commands for video generation and model management.
Sanitization: Absent; the skill does not specify any validation or escaping mechanisms for the user-supplied prompts.
[COMMAND_EXECUTION]: The skill's core functionality relies on executing external command-line tools (comfy-videogen, comfy-models, comfy-tools-setup) using the uv runner. These commands interact with the local file system and manage model downloads.

comfy-minimax-videogen