reclaim-code-entropy skill

A skill that finds redundant abstractions and duplicate state in code and removes them safely, one proved cut at a time.

  • 203 ★ on the repository
  • Last commit
  • MIT
Skill

Extended access

Why this level

  • In apply mode edits and deletes code in the repository.
All reasons and checks

yevanchen/reclaim-code-entropy

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/reclaim-code-entropy

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • In apply mode edits and deletes code in the repository.

This is third-party code. Review the repository files before installing.

What it does

reclaim-code-entropy drives codebase simplification by evidence. The agent first learns what is public, persisted, generated or dependent on dynamics. It then finds redundant representations of one fact, abandoned extension points, forwarding-only wrappers and duplicated lifecycle flags. Before removal the skill requires showing who uses the element, why it was added and what capability would vanish. Audit mode ranks candidates by confidence, cost of error, tradeoff and expected savings, changing nothing. Apply mode makes the one safest cut, preserves observable behavior and runs a proportional check. The result "nothing safe to remove" is valid.

Who it is for. Developers and tech leads who clean up an overgrown codebase with an agent and do not want to break public contracts.

Good fit when

  • A repository accumulated wrappers that only forward calls.
  • Two flags describe the same state transition and you need to see which one is redundant.
  • Before refactoring you need a ranked list of safe candidates without edits.
  • After a feature was dropped, tests and documentation remain only for it.

Not a fit when

  • You need a fast bulk cleanup pass over a static analyzer list.
  • Public exports and data formats must stay compatible and the absence of consumers cannot be proved.
  • The task is adding functionality rather than simplifying.

Example request

Use $reclaim-code-entropy to audit this repository and rank the safest simplification candidates.

Limitations

The skill does not replace the project tests: verification is proportional to the cut and your own test suite confirms final safety. Strings, dynamic dispatch, public exports and migrations cannot be proved by static analysis, so the skill asks for consumer confirmation. It prefers a few proved cuts to a long list of guesses. Installation is described for Codex, and for other agents the folder is copied into their skills directory by hand. The license is MIT. The last commit is dated August 2026.

How to disable. Delete the `reclaim-code-entropy` folder from `~/.codex/skills`. Changes the agent made in the repository are reverted with ordinary git.

FAQ

How do I install the skill?

Ask Codex to install it from the repository path `skills/reclaim-code-entropy` or clone the repository and copy the folder into `~/.codex/skills`. Then start a new task.

How do the modes differ?

Audit ranks candidates and changes nothing. Apply implements the safest cut, preserves behavior and runs a check.

What does the skill look for?

Unused APIs, exports and settings, mirrored state, speculative abstractions, wrappers, duplicated lifecycle and residue of removed implementations in tests and docs.

Can I limit the scope?

Yes. Ask it to check a specific area, for example whether two lifecycle flags are distinct, and the skill examines only that.

What if nothing is found?

That is a valid outcome. The skill does not invent edits to fill a report.

Discussion

No questions or reviews yet.

Skills

A skill that finds AI slop in code: duplicate helpers, empty abstractions, weak tests. It cleans one pattern per commit.

SkillExtended access95Repository stars

A cleanup skill for agent-written code: finds surplus tests, self-checks and fallbacks and deletes them, keeping behavior

SkillExtended access140Repository stars

A skill that runs a full refactor so agents read code more easily: it finds hidden links and needless dependencies.

SkillExtended access1.1KRepository stars

Simplify Codebase

simplify-codebase

A skill that says prove first and delete after: finding accidental complexity in code and safe removal without breakage. The README is in Chinese

SkillExtended access515Repository stars

A code review skill with four independent probes, evidence-backed findings and a merge gate derived from JSON

SkillExtended access479Repository stars

A code review skill across many languages and frameworks: architecture, performance, security, anti-patterns and common bugs

SkillMinimal access2.3KRepository stars13Installs via CLI and agents

README in short

The README describes the problem of accumulating entropy, five work stages, audit and apply modes and sample requests. A Chinese version `README.zh.md` exists.

README badge

Report a problem