Skip to content

ciembor/agent-rules-books / working-effectively-with-legacy-code

working-effectively-with-legacy-code

Verified

Apply Michael Feathers-inspired legacy-code rules when changing hard-to-test code safely with characterization tests, seams, sprout methods, or wrap methods.

working-effectively-with-legacy-code.nano.md

View on GitHub
# OBEY Working Effectively with Legacy Code by Michael Feathers

## When to use

Use when changing poorly tested or poorly understood code under a tight context budget.

## Primary bias to correct

Legacy work starts with control, not cleanup, rewrite, or elegance.

## Decision rules

- Treat code without trustworthy tests as legacy code: state what changes and what must remain.
- Characterize uncertain current behavior before changing it, including ugly behavior consumers may rely on.
- Use the legacy loop: find the change point, find an observation point, create or exploit a seam, break the blocking dependency, test, change, then refactor locally.
- Prefer fast focused tests; use broader harnesses only as temporary first coverage when no narrow test point exists.
- Create the narrowest useful seam for sensing or separation, and break only dependencies that block feedback.
- Use sprout, wrap, parameterize, inject, extract, or override moves when direct edits would be unsafe.
- Keep behavior changes, structural refactorings, and cleanup separate and small.
- Do not leave test-only seams, hidden dependencies, wrappers, globals, subclass tricks, or link/preprocessor tricks without a cleanup plan.

## Trigger rules

- When behavior is unclear, characterize first.
- When constructors, globals, statics, frameworks, I/O, clocks, randomness, environment, or deep object graphs block testing, break one dependency at the narrowest point.
- When a large method or class defeats local reasoning, sketch effects and create a seam before semantic edits.
- When rewrite or broad cleanup feels tempting, choose the next smaller verified move.

## Final checklist

- Behavior characterized?
- Feedback fast enough?
- Dependency isolated?
- One kind of change?
- Safer and clearer now?
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 197kUpdated Claude CodeCodexVerifiedInstall