The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external websites based on user input.\n
Ingestion points: The scripts/parse_douyin.py script extracts URLs from user-provided share text and fetches content from them.\n
Boundary markers: The skill does not use delimiters or instructions to ignore embedded commands within the fetched HTML or metadata.\n
Capability inventory: The skill includes file writing (scripts/download_video.py) and network operations (scripts/parse_douyin.py, scripts/transcribe_audio.py).\n
Sanitization: There is no sanitization of natural language instructions that might be embedded in video titles or descriptions, although filenames are sanitized in scripts/parse_douyin.py.\n- [EXTERNAL_DOWNLOADS]: The skill performs network requests to download video content and access transcription APIs.\n
Fetches data from Douyin domains including v.douyin.com and iesdouyin.com.\n
Communicates with Volcano Engine (Bytedance) at openspeech.bytedance.com for ASR services.\n- [COMMAND_EXECUTION]: The skill uses Python scripts to perform local file system operations.\n
The scripts/download_video.py script writes binary video data to an arbitrary file path provided via command-line arguments.

douyin-downloader