qwen-image-2
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references installation instructions for the
infshCLI tool hosted on the official GitHub repository for inference.sh. These references are neutral and point to the platform's own infrastructure. - [COMMAND_EXECUTION]: The skill uses the
infshcommand-line tool to execute image generation tasks. The use of this tool is restricted via theallowed-toolsconfiguration in the skill metadata, limiting the agent's execution scope to the intended CLI. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user prompts to generate image content. While this creates a surface for indirect instructions, the skill's restricted execution environment (limited to the
infshtool) and the nature of the output (images) mitigate the risk of these instructions affecting the agent's logic or the host system.
Audit Metadata