The Agent Skills Directory

[COMMAND_EXECUTION]: The skill extensively uses Python's subprocess module and Node.js npx to perform media processing tasks.
scripts/generate_tts.py and its backends use ffmpeg for audio concatenation and normalization, and ffprobe for duration validation.
scripts/generate_shorts.py invokes ffmpeg to slice audio segments and npx remotion render to generate vertical videos.
scripts/verify_output.py uses ffprobe to validate the technical specifications of the generated MP4 files.
scripts/learn_design.py uses ffmpeg for frame extraction from reference videos.
scripts/cli.py serves as a dispatcher that spawns the underlying Python scripts as child processes.
[EXTERNAL_DOWNLOADS]: The skill communicates with various established cloud services and manages its own updates via Git.
The TTS backends in scripts/tts/backends/ make authenticated API requests to Microsoft Azure, OpenAI, Google Cloud, Alibaba (CosyVoice), Volcengine (Doubao), and ElevenLabs.
scripts/check_update.sh uses git ls-remote to check for new versions on GitHub and can perform a git pull to update the skill directory, which is gated by explicit user consent in the instructions.
templates/components/LottieAnimation.tsx and MediaSection.tsx support fetching animation data and images from remote URLs (e.g., LottieFiles, Unsplash, Pexels) for use in video compositions.

video-podcast-maker