The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands like curl and yt-dlp using input directly derived from user-provided URLs. The absence of sanitization instructions for these variables creates a risk of command injection if shell metacharacters are included in the input.
[CREDENTIALS_UNSAFE]: The troubleshooting section suggests using the --cookies-from-browser chrome flag with yt-dlp. This instruction allows the agent to access and use sensitive session data from the user's browser, posing a privacy and security risk.
[EXTERNAL_DOWNLOADS]: The skill fetches content from external domains including x.com, twitter.com, t.co, and pbs.twimg.com. While necessary for the skill's functionality, these represent external data sources.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external URL content and metadata. Ingestion points: User-supplied URLs and yt-dlp --dump-json output in SKILL.md. Boundary markers: None present. Capability inventory: curl, yt-dlp, and ls as specified in SKILL.md. Sanitization: No validation or escaping mechanisms are defined.

ljg-x-download