aliyun-wan-video

Pass

Audited by Gen Agent Trust Hub on Apr 30, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXPOSURE]: The scripts in the skill access the sensitive file path ~/.alibabacloud/credentials to retrieve the DASHSCOPE_API_KEY. This is documented as the standard method for authenticating with Alibaba Cloud services and aligns with official vendor recommendations.
  • [EXTERNAL_DOWNLOADS]: The skill relies on the dashscope Python package, which is the official SDK for Alibaba Cloud Model Studio. All external documentation links point to the trusted aliyun.com domain.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided prompts which are interpolated into API requests. This constitutes a standard injection surface for AI agent skills.
  • Ingestion points: User input is ingested via command-line arguments and JSON request files in scripts/generate_video.py and scripts/generate_dancing_video.py.
  • Boundary markers: The skill does not implement specific boundary markers or delimiters when passing user prompts to the DashScope API.
  • Capability inventory: The skill has the capability to perform network requests (via the SDK and urllib.request) and write files to the local system (downloading generated assets).
  • Sanitization: There is no explicit sanitization of prompt strings prior to their use in API payloads.
  • [COMMAND_EXECUTION]: The SKILL.md file contains a validation block that uses py_compile to verify script syntax. This is a standard build-time check and does not involve the execution of untrusted code or commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 30, 2026, 09:12 AM