linkfox-aigc-imagegen-brand-gene-extract
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an input ingestion surface where user-provided brand attributes can influence the generated prompt.
- Ingestion points: User-supplied fields within the
brandKeyobject (such asbrandName,fontStyle) and input images are gathered inSKILL.md(Step 1 and Step 2). - Boundary markers: The prompt construction phase in Step 2 lacks explicit isolation delimiters or structural framing to prevent malicious text inputs from altering the instruction flow.
- Capability inventory: The associated scripts
scripts/onboarding.pyandscripts/save_brand_gene.pypossess capabilities for file writing (open) and network communication (requests,urllib.request). - Sanitization: No data validation or escaping is applied to individual
brandKeyparameters prior to their interpolation into the visual reasoning prompt.
Audit Metadata