douyin-transcribe
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches video media from Douyin's official sharing infrastructure and transcribing models from the ModelScope platform. These operations involve well-known services and are core to the skill's primary functionality.
- [COMMAND_EXECUTION]: Executes curl and ffmpeg via subprocess.run using argument lists, which prevents shell injection vulnerabilities.
- [INDIRECT_PROMPT_INJECTION]: Processes external video titles and URLs (ingestion points) into a markdown output body (boundaries). Includes filename sanitization to prevent directory traversal when utilizing file-write capabilities.
- [DYNAMIC_EXECUTION]: References the trust_remote_code parameter for loading models from the ModelScope repository, which is a standard procedure within the FunASR and ModelScope ecosystems for specialized models.
Audit Metadata