qwen-image-2-pro
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: No security issues were detected. The skill follows best practice by scoping tool access to the specific 'belt' CLI in its frontmatter and provides clear documentation for its intended use. All referenced URLs are related to the official service or provide standard examples.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes external prompts for image generation. However, this is inherent to its primary function and managed through standard JSON input structures.
- Ingestion points: 'prompt' and 'reference_images' inputs in SKILL.md.
- Boundary markers: Inputs are delimited within JSON objects.
- Capability inventory: Uses 'Bash(belt *)' to run inference applications.
- Sanitization: Content is passed as data to the generation model via the CLI tool.
- [EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install the platform's CLI tool and mentions dependencies for the Python SDK.
- Sources: 'belt-sh/cli' via npx and 'inferencesh' via pip.
Audit Metadata