The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local files, creating a minor surface for indirect prompt injection.
Ingestion points: The scripts analyze_video.py and analyze_drama.py read local video files, script.txt, and drama_info.json to generate analysis reports.
Boundary markers: There are no explicit delimiters or instructions provided to the model to ignore potentially malicious content embedded within the processed text or metadata.
Capability inventory: The skill possesses the capability to read from and write to the local filesystem and perform network operations to the Volcengine API.
Sanitization: No sanitization or escaping is performed on the content of external files before they are interpolated into the final Markdown report.
[DYNAMIC_EXECUTION]: The skill uses __import__('datetime') within the script files (analyze_video.py, analyze_drama.py) to generate timestamps for the output reports. This is a dynamic import of a standard library module from a static string literal and is a benign programming practice.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to the Volcengine Ark API (ark.cn-beijing.volces.com) for its core functionality. This is a well-known cloud service provided by ByteDance and represents normal operation for a video understanding tool.

volcengine-video-understanding