alibabacloud-video-prompt-architect
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as a specialized text transformer for AI prompt engineering. It lacks any functionality for command execution, network requests, or sensitive file access. All operations are confined to the text processing context of the LLM.- [SAFE]: The workflow consists of intent analysis, structured decomposition, and model-specific adaptation performed within the LLM's context. No external tools, libraries, or scripts are invoked or requested.- [SAFE]: The skill includes quality-focused features like a physical plausibility check, which identifies and corrects unrealistic descriptions (e.g., gravity-defying actions) and provides mandatory user notifications, acting as a functional guardrail.- [SAFE]: Strict output formatting rules are implemented to ensure that final prompts are delivered in isolated code blocks, with a absolute prohibition on additional text after the final delivery to prevent output confusion or extraneous generation.
Audit Metadata