How does this differ from the /decomp skill?

While /decomp focuses on matching assembly and reducing diff percentages, /understand is specifically for improving code readability through naming and documentation.

Can I use this for other GameCube projects?

This skill is specifically optimized for the Super Smash Bros. Melee repository structure, module prefixes (ft_, gr_, lb_), and specific engine patterns.

How do I handle unknown struct fields?

Use the convention unk or x until the field's purpose is verified, at which point it should be renamed to descriptive snake_case.

What are subdirectory worktrees?

They are isolated directories for specific modules (like fighters or stages) that allow agents to work in parallel without causing git conflicts in the main directory.

What documentation tags should I use?

The skill supports standard Doxygen-style tags like @brief, @param, @return, @at for offsets, and @sz for field sizes.

Melee Code Understanding

Name: Melee Code Understanding
Author: itsgrimetime

byitsgrimetime

•

Game Development

Documents and identifies function purposes, struct fields, and data types within the Super Smash Bros. Melee decompilation project.

About

This skill provides an expert reverse-engineering assistant for the Melee decompilation project, helping developers translate address-based functions and unknown memory fields into human-readable code. It streamlines the workflow of documenting module logic, identifying game mechanics like fighter state machines or stage features, and maintaining consistency across the codebase using specialized subdirectory worktrees and context-aware analysis.

Key Features

Subdirectory-based worktree management for parallel developer collaboration
Cross-reference gathering to trace function callers and callees
3 GitHub stars
Automated documentation generation for @brief, @param, and @return tags
Descriptive naming for address-based functions and unknown variables
Struct field discovery and offset-based memory analysis

Use Cases

Mapping out unknown fighter data structures and action state variables
Documenting complex stage collision and item behavior logic for better readability
Converting address-based function labels into descriptive, logic-based names

About

Key Features

Subdirectory-based worktree management for parallel developer collaboration
Cross-reference gathering to trace function callers and callees
3 GitHub stars
Automated documentation generation for @brief, @param, and @return tags
Descriptive naming for address-based functions and unknown variables
Struct field discovery and offset-based memory analysis

Use Cases

Mapping out unknown fighter data structures and action state variables
Documenting complex stage collision and item behavior logic for better readability
Converting address-based function labels into descriptive, logic-based names