The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements a comprehensive research pipeline that ingests data from external sources (web search, transcripts, and writings). While this represents an indirect prompt injection surface, the risk is inherent to the skill's primary function of information synthesis and is mitigated by the structured verification phases.
[COMMAND_EXECUTION]: The instructions describe the use of sub-agents and external tools like yt-dlp for data collection. These actions are within the intended operational scope of the skill for media processing and file management.
[EXTERNAL_DOWNLOADS]: The skill facilitates the fetching of research materials from public repositories and media platforms. The sources mentioned are consistent with the objective of gathering public intellectual property for cognitive modeling.

huashu-nuwa