The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes data from external untrusted sources, including Feishu documents, Word files, and Markdown content. This creates an indirect prompt injection surface where malicious instructions embedded in the ingested documents could potentially influence the AI's content adaptation or distribution behavior. No specific boundary markers or sanitization logic are mentioned to mitigate this risk.\n- [EXTERNAL_DOWNLOADS]: The skill directs users to download a ZIP file from ziliu.online/extension and manually install it in developer mode. While this is a vendor-owned resource necessary for the primary function of the skill (bypassing API limitations of platforms like WeChat or Little Red Book), manual installation of external browser extensions carries standard software supply chain risks.\n- [COMMAND_EXECUTION]: The documentation provides an example Python code snippet for browser automation. This is presented as an integration template rather than an automated script executed by the skill itself, but it involves interacting with local browser environments.

ziliu-publisher