What problem does it solve?
This Skill guides an agent to reliably extract a chess position from a photographic or scanned board image, validate the detected position for legality, and determine the optimal or forced moves using a chess engine without guessing when image detection is unreliable.
Core Features & Use Cases
- Calibrate image-specific color signatures for board squares and piece colors to minimize false detections.
- Detect an 8x8 grid, determine square occupancy, and produce a human-readable debug representation for incremental verification.
- Validate detected positions with strict rules (exactly one king per side, pawn and piece count limits, total pieces ≤32) and decline to proceed on invalid detections.
- Load validated positions into a UCI chess engine (e.g., Stockfish via python-chess) to compute best moves, verify legality, and confirm mate sequences.
- Use case: turn a user-supplied photo of a tournament board into a verified move suggestion or request FEN/clarification when detection fails.
Quick Start
Provide the board image and ask the assistant to detect pieces, validate the position, and return the best legal move with verification.