NYC

bilibili-analyzer

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The tool is designed to ingest and analyze untrusted external data (video frames and metadata) from Bilibili.
  • Ingestion points: Video URLs processed by main.py.
  • Boundary markers: None explicitly documented in the provided files.
  • Capability inventory: Downloads video files via yt-dlp, executes system commands via ffmpeg, and sends data to an AI for analysis.
  • Sanitization: Documentation does not specify sanitization methods for ingested metadata before analysis.
  • [Command Execution] (LOW): The tool's primary function involves executing local scripts (main.py) and calling the system utility ffmpeg to process video files. This is consistent with its stated purpose.
  • [External Downloads] (LOW): Installation instructions guide the user to download standard, reputable Python packages (yt-dlp, requests) and system binaries (ffmpeg) from official repositories.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:06 PM