sketch-image-prompt
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions for the AI to perform text analysis and creative writing. It does not include any executable scripts or access to sensitive system resources.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text to generate summaries and image prompts, which creates an indirect prompt injection surface. However, the available tools are limited to UI visualization widgets, and there is no mechanism for data exfiltration or system-level exploitation. Ingestion points: User-provided text content. Boundary markers: Absent. Capability inventory: Restricted to UI visualization tools (visualize:show_widget, visualize:read_me). Sanitization: Absent.
- [NO_CODE]: No executable scripts are included in the skill. The logic is implemented through natural language instructions for the AI model.
Audit Metadata