p-image
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides links to configuration and installation files hosted on the author's official GitHub repository (
inference-sh/skills). These references are neutral and point to the vendor's own infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits a potential surface for indirect prompt injection as it processes user-supplied text for image generation.
- Ingestion points: User-provided image generation prompts are accepted in the
--inputparameter of thebeltCLI examples inSKILL.md. - Boundary markers: Prompts are encapsulated within JSON objects passed to the CLI, providing structural separation from the command logic.
- Capability inventory: The skill is limited to invoking the
beltCLI tool for remote image generation via Bash. - Sanitization: No specific sanitization is mentioned in the prompt instructions, relying on backend model safety guardrails.
- [SAFE]: All referenced domains (
inference.sh), tools (belt), and repositories are owned by the verified authorinference-sh, representing legitimate vendor functionality.
Audit Metadata