The Agent Skills Directory

[COMMAND_EXECUTION]: The skill orchestrates multiple shell commands including directory creation, navigation, and execution of Python scripts and CLI tools. It relies on user-defined shell aliases to perform video downloads and transcriptions. The list_channel_videos.py script uses the subprocess module to call yt-dlp using an argument list, which is a standard method for executing external processes.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from an external source (YouTube video metadata) which is then used to drive agent actions and shell commands.\n
Ingestion points: The list_channel_videos.py script fetches video titles, IDs, and URLs from a specified YouTube channel.\n
Boundary markers: The SKILL.md file provides a structured 'Execution Checklist' that guides the agent through the process, separating the parsing of external data from the execution of processing steps.\n
Capability inventory: The skill has the capability to execute shell commands, manage files and directories, and call external CLI tools for media processing.\n
Sanitization: The instructions include a mandatory sanitization routine for video titles (replacing illegal characters, removing non-alphanumeric characters, and truncating) before the titles are used in directory creation or file paths.

batch-processor