The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands and external utilities such as yt-dlp and jq for media processing. It further provides detailed configurations for establishing automated persistence through crontab and systemd user timers to handle recurring archive updates.
[DATA_EXFILTRATION]: The skill facilitates network communication with external media platforms including YouTube, Vimeo, and Twitch to retrieve media files and metadata. These operations target well-known services and do not involve sensitive local data.
[PROMPT_INJECTION]: The skill processes untrusted metadata from remote media platforms, which constitutes an indirect prompt injection surface.
Ingestion points: Untrusted data enters the agent context via info.json metadata generated by yt-dlp during playlist and channel extraction (processed in MediaOrganizer and MediaLibrary classes).
Boundary markers: No specific boundary markers or instructions to ignore embedded content are provided for the agent when processing this data.
Capability inventory: The skill possesses capabilities for file system modification (shutil.move, os.makedirs), shell command execution (subprocess.run), and database writes (sqlite3).
Sanitization: The MediaOrganizer class implements a _sanitize method that uses regex to strip common directory traversal and illegal characters from metadata before using them in file paths.

playlist-archiver