The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module to invoke local binaries including ffmpeg, ffprobe, and yt-dlp for media processing. All commands are executed with list-based arguments to prevent shell injection, and the operations are strictly aligned with the skill's primary purpose.
[EXTERNAL_DOWNLOADS]: The skill downloads video metadata, subtitles, and media files from user-provided URLs using yt-dlp. It also performs audio transcription by sending data to well-known providers (Groq and OpenAI) only when necessary and with explicit user configuration.
[CREDENTIALS_UNSAFE]: API keys for transcription services are managed securely. The skill scaffolds a .env file in the user's home directory (~/.config/watch/.env) and explicitly sets file permissions to 0600 (read/write by owner only), which is the standard best practice for local secret storage.
[SAFE]: The skill documentation and metadata are consistent with its actual implementation. It provides clear setup instructions and does not employ any obfuscation, persistence, or privilege escalation techniques.

watch