interior-design-expert
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- PROMPT_INJECTION (LOW): The skill identifies a surface for indirect prompt injection in the 'AI Visualization Prompts' section of SKILL.md. User-provided room parameters and style descriptions are interpolated into templates for the Stability AI and Ideogram tools without explicit sanitization or boundary delimiters.\n
- Ingestion points: User input for Style, Room Type, Key Features, and Mood in SKILL.md prompt templates.\n
- Boundary markers: Absent; the templates directly concatenate user-controlled strings.\n
- Capability inventory: Uses mcp__stability-ai__stability-ai-generate-image and mcp__ideogram__generate_image for external tool execution.\n
- Sanitization: None detected; the skill relies on the agent to construct the final prompt based on untrusted user descriptions.- EXTERNAL_DOWNLOADS (SAFE): The Python code snippets in the references directory (space-planning.md) import external libraries.\n
- Evidence: Imports for 'ortools' and 'numpy'.\n
- Trust Status: 'ortools' is maintained by Google, and 'numpy' is a core scientific package; both originate from trusted sources per [TRUST-SCOPE-RULE].
Audit Metadata