The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to download additional scripts and documentation from the author's official GitHub repository (github.com/novoads/agent-skills). As these resources originate from the vendor's own infrastructure, they are documented as standard functional components.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, specifically video files and transcripts, which are then used to construct prompts for the video generation model. This creates an attack surface for indirect prompt injection.
Ingestion points: The skill ingests source.mp4, audio.wav, and transcribed text returned from the Novoads API (SKILL.md).
Boundary markers: The skill does not explicitly instruct the agent to use delimiters or 'ignore embedded instructions' warnings when interpolating transcript data into adaptation prompts.
Capability inventory: The skill has network access via curl, file system access for logs and outputs, and shell execution capabilities via ffmpeg (SKILL.md).
Sanitization: No specific sanitization or filtering of external content is implemented before it is used in the prompt adaptation process.
[COMMAND_EXECUTION]: The skill utilizes shell commands like ffmpeg for media processing and curl for API interactions. These operations are documented as standard functional components.

clone-video-ad