Reduce codebase entropy
Review a change or area with a bias toward deletion, locality, and fewer concepts.
Ready to paste
Replace the bracketed context, then use it in your agent.
Act as a senior engineer optimizing for a smaller and more truthful codebase. Code looks organized but requires too many files, helpers, layers, or concepts to understand a simple behavior. Context to use: - Target area or diff: [reference] Process: 1. Explain the behavior in plain domain language. 2. Count the concepts, hops, and abstractions required to follow it. 3. Identify helpers or layers that do not hide meaningful complexity. 4. Move behavior closer to the data and event that give it meaning. 5. Compare total code and cognitive load before and after. Constraints: - Preserve behavior. - Do not centralize unrelated code merely to reduce file count. - Keep abstractions that enforce a real invariant or isolate volatility. Return: - Complexity diagnosis - Candidates for deletion or inlining - Simpler structure - Behavior-preservation checks
Use when
Code looks organized but requires too many files, helpers, layers, or concepts to understand a simple behavior.
Expected return
- — Complexity diagnosis
- — Candidates for deletion or inlining
- — Simpler structure
- — Behavior-preservation checks
Related prompts
Map an unfamiliar codebase
Build a useful model of modules, data flow, boundaries, and domain language before changing code.
Find architecture opportunities worth taking
Identify improvements that reduce future change cost instead of merely rearranging files.
Design a deep module
Place substantial behavior behind a small, stable interface at a real domain seam.