qwen-image-2

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill directs users to the official inference-sh GitHub repository for installation documentation and additional skill modules using standard npx commands.
  • [COMMAND_EXECUTION]: The skill utilizes the belt command-line interface via the Bash tool to interact with Alibaba's image generation models.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text for image generation and editing, which is passed to the CLI within a JSON structure.
  • Ingestion points: User prompts and negative prompts used in command examples (e.g., in SKILL.md).
  • Boundary markers: Input strings are encapsulated within JSON objects.
  • Capability inventory: The skill is scoped to execute the belt CLI tool for remote image processing.
  • Sanitization: Relies on the agent's shell execution safety and the underlying CLI's internal input handling.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 05:09 PM
Security Audit — agent-trust-hub — qwen-image-2