The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of CLI tools (yt-dlp, ffmpeg) which involves passing user-provided inputs such as URLs and filenames directly to shell commands. If the agent does not properly sanitize these inputs, it could lead to command injection vulnerability.
[PROMPT_INJECTION]: There is a potential for indirect prompt injection. The skill is designed to ingest data from external sources (video sharing platforms) via yt-dlp.
Ingestion points: Video titles, descriptions, and other metadata fetched from remote URLs during the download process.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are provided for the fetched metadata.
Capability inventory: The skill utilizes subprocess execution for media manipulation and downloading.
Sanitization: There are no sanitization steps defined for the metadata retrieved from external sites.
[EXTERNAL_DOWNLOADS]: The skill references the installation and update of yt-dlp using Python's package manager. Since yt-dlp is a well-known and widely used tool for media retrieval, this reference is documented neutrally.

c-video