pptx-posters

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill’s PPTX/HTML workflow ingests user-provided free text prompts via scripts/generate_schematic.py → generate_schematic_ai.py (prompt is sent as messages to the Gemini/OpenRouter review and generation calls), meaning outsider-authored text can directly influence runtime LLM processing.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The scripts (scripts/generate_schematic_ai.py) make runtime API calls to https://openrouter.ai/api/v1 whose text/image responses are used to generate images and whose review critique is injected into improved prompts, so external content directly controls prompts and execution.