soia-pkm-clip-douyin

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill requires the installation of the playwright library and its browser binaries to intercept signed API responses from Douyin. It also downloads video media (MP4) from Douyin's infrastructure using the Python standard library urllib. All external resources and dependencies originate from well-known, legitimate services.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from Douyin video descriptions and author metadata, which is subsequently stored in the user's local vault.\n
  • Ingestion points: The desc and author nickname fields in the aweme JSON object intercepted by scripts/archive_douyin.py.\n
  • Boundary markers: The ingested content is isolated within a ## 原文 (Original Text) header in the generated Markdown file.\n
  • Capability inventory: The skill possesses the ability to write to the filesystem (vault and downloads folder) and perform network requests to Douyin domains.\n
  • Sanitization: Filenames are sanitized for filesystem compatibility, but the video description text is written to the vault without additional escaping or sanitization against potential command injection targeting downstream agent actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 08:05 AM
Security Audit — agent-trust-hub — soia-pkm-clip-douyin