short-drama-video-prompts
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows secure development practices for AI agents. It focuses on local processing of drama project files (scripts, storyboards, and visual settings) to generate prompts for specific video models like MiniMax H3 and Seedance.
- [INDIRECT_PROMPT_INJECTION]: The skill inherently handles untrusted external data (drama scripts and user-provided descriptions). However, it includes robust guidelines for using delimiters, such as
<d>tags and quotation fences, to isolate verbatim dialogue from instructions. This practice significantly reduces the risk of the model misinterpreting data as new commands. - [COMMAND_EXECUTION]: The skill includes several Python scripts (
scripts/selftest.py,scripts/container_check.py, etc.) used for validating the mathematical and structural integrity of local JSONL files. These scripts utilize standard libraries for JSON parsing and arithmetic, performing no network operations or sensitive file system access beyond the project directory. - [DATA_EXPOSURE]: There is no evidence of hardcoded credentials, access to sensitive system paths (e.g.,
.ssh,.aws), or network calls to external domains. All file operations are confined to the defined drama project structure.
Audit Metadata