bilibili-render-pdf

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions specify the use of command-line tools for functional tasks, including yt-dlp for video and subtitle retrieval and whisper for audio transcription.
  • [EXTERNAL_DOWNLOADS]: The skill fetches video metadata, thumbnails, and subtitles from Bilibili domains (bilibili.com and b23.tv). These are legitimate external resources required for the primary function.
  • [PROMPT_INJECTION]: The skill processes untrusted external content (Bilibili subtitles and Whisper-generated transcripts) and incorporates it into a LaTeX document that is subsequently compiled. This creates a surface for indirect prompt injection.
  • Ingestion points: Bilibili subtitles and transcription files (SKILL.md).
  • Boundary markers: None explicitly mentioned for the subtitle content within the LaTeX template.
  • Capability inventory: Shell access for yt-dlp, whisper, and LaTeX compilation.
  • Sanitization: No specific filtering or sanitization rules for subtitle content are mentioned before insertion into the .tex file.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 06:01 AM