What problem does it solve? Site inspectors need a consistent, objective way to rate the cleanliness of plantation staff quarters from photos without manually scoring each surface or mixing up cleanliness with structural condition. ## Core Features & Use Cases - Per-Surface Cleanliness Rating: Rates floor, walls, bedding, and surfaces on a good/moderate/poor scale using a vision model, returning structured JSON output. - Multi-Photo Handling: Analyzes multiple photos separately and takes the worst rating per surface, noting which photo showed each issue. - Clear Scope Boundaries: Focuses only on cleanliness — it does not count furniture, assess structural condition, or store reports, making it safe to run as one of three parallel inspection checks. - Use Case: A plantation supervisor photographs a staff quarters room and asks for a cleanliness rating; the skill returns per-surface ratings plus an overall summary like "Floor needs sweeping, bedding needs changing — overall moderate cleanliness." ## Quick Start Send a photo of the staff quarters and ask the agent to check the cleanliness of this room.
