The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches audio assets from well-known services to support video creation.
Downloads generated voiceover audio from the ElevenLabs API (api.elevenlabs.io).
Downloads royalty-free background music from established media providers such as bensound.com and pixabay.com.
[COMMAND_EXECUTION]: The skill uses system tools to process media and manage development environments.
Executes ffmpeg and ffprobe via Python's subprocess module in scripts/generate_voiceover.py for audio analysis and processing.
Uses shell commands like git log, head, and ls in SKILL.md to analyze the local repository and provide video context.
Utilizes npm and npx for installing dependencies (lucide-react) and executing Remotion video rendering tasks.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes project data to generate scripts, which represents a potential surface for processing untrusted content.
Ingestion points: The skill reads local repository data including git log and project source files (e.g., README.md, app/models/) to generate promotional content.
Boundary markers: The workflow relies on interactive selection menus and a manual preview phase where the user must approve the generated video before final rendering.
Capability inventory: The skill has the ability to write files to the local disk (audio and video artifacts) and perform network requests to the ElevenLabs API.
Sanitization: While no explicit sanitization logic is shown in the scripts, the architecture incorporates human-in-the-loop verification at critical stages of the creative process.

promo-video