About this project

Ikemen Combo Finder is a Python-based tool designed to automatically discover possible combo chains for characters in Ikemen GO and MUGEN fighting games. It parses character definition files (.def, .cmd, .cns/.st), constructs a transition graph of valid move sequences based on heuristic rules (e.g., Light → Medium → Heavy, Normal → Special → Super), then uses a genetic algorithm to evolve optimal combos. These are validated via a rule-based simulator that estimates damage scaling and hitstun. The project is openly acknowledged as experimental — developed largely with AI assistance by a non-programmer — and contains known flaws, inconsistent updates, and limited support. Output includes JSON files listing extracted moves and discovered combos. While useful for exploration or prototyping, results should not be treated as authoritative due to simplified modeling of game mechanics. Users must manually configure character paths in config.yaml and run extraction/search scripts. The codebase is messy and not engineered for production; developers are encouraged to fork or rewrite rather than expect maintenance. Licensed under MIT.