The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of yt-dlp via pip. yt-dlp is a well-known, legitimate, and widely-used open-source utility for interacting with social media media content.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection because it ingests and processes data from external social media platforms controlled by third parties.
Ingestion points: Blogger profiles, bios, descriptions, and post content from Douyin, Xiaohongshu, and Bilibili are retrieved using the browser tool (defined in SKILL.md).
Boundary markers: The instructions do not define any delimiters or explicit safety warnings to the agent to disregard instructions that might be embedded within the crawled profile information or video metadata.
Capability inventory: The agent uses browser for web navigation, yt-dlp for downloading content, and ffmpeg for processing media files.
Sanitization: There is no evidence of sanitization, validation, or escaping of the retrieved external data before it is interpolated into the agent's context for AI analysis.

influencer-analyzer