modellix-image-prompt-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, such as prompt injection, obfuscation, or data exfiltration, were detected. The skill instructions focus entirely on visual art direction and prompt optimization.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user visual briefs (SKILL.md). The instructions mitigate potential injection by requiring the agent to use quotation marks for exact copy and prohibiting the fabrication of specifications. The skill operates without external tools, limiting the potential impact of adversarial input.
- [EXTERNAL_DOWNLOADS]: The skill mentions an external GitHub repository for attribution (alchaincyf/huashu-gpt-image) and suggests using official models from Google (google/nano-banana-2-lite). These are neutral references to external sources and do not involve automated execution or downloads.
Audit Metadata