The Agent Skills Directory

[SAFE]: No security issues were detected. The skill implements robust security controls for handling private user data, including explicit per-turn authorization requirements and strict read-only boundaries that forbid actions like liking, commenting, or following. It avoids data exfiltration by saving bookmarks only to local files and prohibiting the display of private content or sensitive URLs in the chat interface. Technical operations such as local command execution (via the Field Theory CLI) and browser-based data extraction are used strictly for the skill's primary purpose and are implemented with safe, controlled logic. The processing of external social media data is mitigated by specific ID extraction and validation using regular expressions, which prevents indirect prompt injection risks by avoiding the direct interpolation of raw platform strings into the agent's context.

yichen-bookmarks-export