douyin-transcript-exporter
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from Douyin video titles, descriptions, and transcripts to generate AI-based summaries and categories. This creates a surface for indirect prompt injection if the source content contains malicious instructions intended to manipulate the AI's output. * Ingestion points: Video content and transcripts are scraped from external Douyin URLs in SKILL.md (Step 2 and 3) and references/douyin.md. * Boundary markers: The instructions do not define specific delimiters or 'ignore' instructions for the AI processing step (Step 3.5). * Capability inventory: The skill possesses the capability to write records to Feishu Base via lark-cli and control the browser environment via computer_use_tool. * Sanitization: No sanitization or filtering of the fetched transcript text is described prior to AI processing.
- [COMMAND_EXECUTION]: The skill implements a fallback path that executes a local Python script to process external video URLs. * Evidence: SKILL.md (Step 3) and references/douyin.md describe executing 'python3 scripts/minutes/social_video_to_minutes.py' using the target video URL as an argument.
Audit Metadata