logo-design-guide
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides a link to installation instructions hosted at a GitHub repository for the inference.sh service to set up the necessary CLI tools.
- [PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it facilitates the processing of user-supplied natural language prompts for image generation.
- Ingestion points: User-provided strings within the
promptfield of the JSON input forbelt app runcommands. - Boundary markers: Prompts are encapsulated within structured JSON objects passed as command arguments, which provides basic delimitation.
- Capability inventory: Uses the
beltCLI to invoke remote AI models for image generation, refining, and upscaling. - Sanitization: No explicit sanitization of the prompt strings is performed by the skill instructions, relying on the target AI model's internal safety guardrails.
Audit Metadata