qwen-image-lite
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided prompts and image URLs to perform editing tasks. Ingestion points include the
promptandimage_urlfields. Capability inventory includes file system write access (~/.pixazo/api-key) and network requests togateway.pixazo.ai. No specific boundary markers or sanitization methods are provided for the user-supplied data. - [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands to manage credentials, such as using
chmod 600on the API key file. It also provides a command usingnpxto fetch additional skills from the vendor's repository. - [EXTERNAL_DOWNLOADS]: The skill fetches model reference metadata from the vendor's domain (
pixazo.ai) to ensure correct API parameter usage.
Audit Metadata