The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted product metadata and audience descriptions which are then interpolated into image generation prompts. This creates a surface for indirect prompt injection where malicious data could influence the image generation tool. 1. Ingestion points: The skill receives data through the generate_garment_lifestyle_photo tool schema, specifically the garment_type and target_audience properties. 2. Boundary markers: The skill uses a structured prompt template with clear section headers like SCENE & CONTEXT and GARMENT (HERO PIECE) which provide some structural delimitation. 3. Capability inventory: The skill utilizes an imageGenTool and accesses a remote garment_image URL. 4. Sanitization: There are no instructions for sanitizing, escaping, or validating the input data before it is included in the prompt construction.

garment-lifestyle-photography