The Agent Skills Directory

[COMMAND_EXECUTION]: Uses curl for API requests and ffmpeg for video processing. These are legitimate uses of the tools specified in the allowed-tools configuration and are necessary for the skill's media generation workflow.
[EXTERNAL_DOWNLOADS]: Interfaces with api.varg.ai and mcp.varg.ai. These domains are used to facilitate the core image and video generation features of the skill.
[CREDENTIALS_UNSAFE]: Employs the VARG_API_KEY environment variable for API authentication, which is a standard and secure practice for secret management.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided script data as part of its generation workflow.
Ingestion points: User-supplied dialogue in script.md and character reference URLs.
Boundary markers: None identified; dialogue lines are interpolated directly into video generation prompts.
Capability inventory: Network access via curl, file system operations (Read/Write/Edit), and media processing with ffmpeg.
Sanitization: No input sanitization or validation is implemented for the processed text.

varg-podcast-video