h3-prompt-writing
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill and its referenced files contain only formatting instructions and structural templates for video generation prompts. There are no executable scripts, binaries, or commands present.
- [INDIRECT_PROMPT_INJECTION]: The skill defines workflows for processing user-supplied dialogue, lyrics, and scene text. It implements boundary markers by instructing the agent to wrap this content in
<d>...</d>tags (e.g.,<d>[English] exact content</d>). This use of specific delimiters for untrusted data is a standard practice and, combined with the lack of tool-calling capabilities, presents no significant risk.
Audit Metadata