linkfox-aigc-imagegen-brand-gene-extract
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. Untrusted data enters the agent context via user-provided images and brandKey parameters in SKILL.md. This data is interpolated into a visual reasoning prompt in Step 2 without boundary markers or instructions to ignore embedded content. The skill possesses capabilities for subprocess execution and file writes across its scripts, and no sanitization of the input parameters is performed before prompt assembly.
- [COMMAND_EXECUTION]: The skill executes shell commands to invoke the save_brand_gene.py script and interact with the linkfox-aigc-textgen skill. These operations are used for managing session-specific data files and performing visual analysis.
- [EXTERNAL_DOWNLOADS]: Includes a script function to download media from remote URLs for processing within the local session. The skill also references a vendor-hosted ZIP file for downloading onboarding utilities, which is treated as a standard vendor resource.
Audit Metadata