The Agent Skills Directory

[SAFE]: No malicious patterns or security violations were detected. The skill's behavior is consistent with its stated purpose of browser-based social media automation.- [COMMAND_EXECUTION]: The skill utilizes shell commands such as ls, cat, file, and cp to verify the presence, format, and size of local media files and to move them to the upload directory. These operations are performed on user-specified paths for legitimate asset management.- [DATA_EXFILTRATION]: The skill captures screenshots of the browser interface, including login QR codes, and sends them to the user. This is a functional requirement for handling platform authentication and provides transparency into the automation process.- [PROMPT_INJECTION]: The skill processes external text files for content publishing. While this presents an indirect prompt injection surface, the risk is mitigated by a predefined workflow and a 'Human-in-the-Loop' policy that requires manual confirmation before any content is published.

douyin-publisher