Compare behavior-preserving cleanup workflows by scope, safety checks, language assumptions, and willingness to change architecture.
The job
Includedcode simplification · behavior-preserving structural cleanup · maintainability improvements with verification
Not this pagebug diagnosis · feature implementation · large architecture migration
30-second route
It is the closest explicit match to the task and supplies behavior-preservation, scope, and verification guidance without a declared host-tool dependency.
Editorial confidence: highIts workflow is purpose-built to review recent diffs with repository evidence, risk tiers, and targeted verification.
Editorial confidence: highIt provides the most structured codemap-to-plan-to-verified-execution process among the selected options.
Editorial confidence: mediumThe supplied workflow explicitly depends on host-specific tools.
Editorial confidence: highBefore you install
Only the selected Refactor and Simplify Code workflows make operational assumptions about host capabilities; no supplied source provides a universal installation procedure for these selected Skills.
Read the intervention guide →Source-led comparison
Facts are extracted from each selected Skill’s declared workflow. Signals describe public adoption; they are not a universal score.
Source notes
Why these candidatesEach candidate comes from tracked public sources and has a source-attributed adoption signal. This is a shortlist, not a catalogue or global ranking.
What this cannot establishWe read declared instructions and public signals. We do not run skills, benchmark outputs, audit security, or claim a universal best choice.
Use it forGeneral behavior-preserving simplification of working code for clarity and maintainability.
Avoid ifThe code is already clear, not yet understood, performance-critical where simplicity may slow it down, or being replaced wholesale.
ActionUse as the default cleanup workflow: understand intent, make one simplification at a time, and verify after each change.
EditsTargets nesting, long functions, unclear names, duplication, dead code, redundant abstractions, and unnecessary assertions.
SetupNo setup or invocation is declared in the supplied source.
ContextLanguage-agnostic workflow with TypeScript, Python, and React/JSX examples.
OutputIncremental, reviewable simplifications with a final verification checklist.
ControlConservative: preserve outputs, side effects, error behavior, ordering, project conventions, and scope.
InvocationNot yet mapped from the source.
Install scopeNot yet mapped from the source.
Agent fitHigh: its stated purpose exactly matches working code that needs clarity without behavior changes.
Watch outDo not equate fewer lines with clearer code; retain meaningful abstractions and error handling.
Explicitly applies when code works but is harder to read, maintain, or extend, without changing behavior.SKILL.md:lines 1-3 · inspected 2026-07-20 · high confidence
Defines simplification as reducing complexity while preserving exact behavior.SKILL.md:lines 10-12 · inspected 2026-07-20 · high confidence
Provides a final verification checklist covering tests, build, lint, conventions, and error handling.SKILL.md:lines 319-331 · inspected 2026-07-20 · high confidence
Use it forA bounded cleanup pass on uncommitted, staged, committed, branch, or explicitly named recent changes.
Avoid ifThere is no identifiable changed code, the diff is very large and cannot be scoped, or three parallel reviewers are unavailable or too costly.
ActionCapture the relevant diff, run reuse, quality, and efficiency review in parallel, then apply only safe or verified careful fixes.
EditsFinds duplicate implementations, redundant state, weak abstractions, unnecessary work, and silent failures; excludes risky contract-changing edits from automatic application.
SetupRequires an environment that supports its documented `delegate_task` batch workflow for three concurrent reviewers.
ContextDiff-first and repository-aware; each reviewer receives the full diff and searches the surrounding codebase.
OutputStructured findings with confidence and SAFE/CAREFUL/RISKY tiers, followed by targeted verification and a summary.
ControlRisk-tiered: auto-apply only proven-safe cleanup; verify careful edits; flag risky changes for human review.
InvocationNot yet mapped from the source.
Install scopeNot yet mapped from the source.
Agent fitHigh for recent-change cleanup, with unusually explicit evidence and risk handling.
Watch outIt intentionally limits cleanup to the diff and minimal surrounding edits; it is not a module-wide rewrite workflow.
Describes a parallel three-agent cleanup of recent changes and aggregation before applying worthwhile fixes.SKILL.md:lines 14-22 · inspected 2026-07-20 · high confidence
Supports explicit diff scopes and stops when no changed code can be identified.SKILL.md:lines 47-74 · inspected 2026-07-20 · high confidence
Requires evidence searching, confidence/risk reporting, and prohibits auto-application of risky contract-affecting changes.SKILL.md:lines 88-105 · inspected 2026-07-20 · high confidence
Use it forA larger, explicitly scoped file/module/project refactor when you want codemapping, test assessment, a detailed plan, and continuous verification.
Avoid ifThe host cannot supply its documented OMO agents, LSP, AST-grep, todo, and test tooling; use a less tool-bound workflow instead.
ActionInvoke `/refactor` with a target, scope, and safe strategy; clarify ambiguous goals before mapping dependencies and making verified steps.
EditsCovers renames, structural changes, pattern transformations, and modernization while assessing impact and test coverage.
SetupIts workflow explicitly assumes `call_omo_agent`, LSP tools, AST-grep, TodoWrite, and the repository’s test/type-check tooling.
ContextGeneral codebase-aware refactoring workflow, with safe as the default strategy and broader aggressive changes optional.
OutputIntent confirmation, codemap, verification plan, atomic refactoring steps, and final regression checks.
ControlProcess-heavy and safety-oriented: pauses for unclear goals or low coverage, verifies after every step, and stops on failures.
Invocation/refactor <refactoring-target> [--scope=<file|module|project>] [--strategy=<safe|aggressive>]
Install scopeNot yet mapped from the source.
Agent fitHigh for a deliberate multi-file refactor, but less portable than the first two recommendations.
Watch outThe source’s command examples and orchestration are specific to its host environment, so adaptation may be required.
Explicitly triggers on refactor, cleanup, restructure, extract, simplify, and modernize requests.SKILL.md:lines 1-3 · inspected 2026-07-20 · high confidence
States codebase mapping, risk assessment, planning, precise execution, and tests after each change.SKILL.md:lines 30-39 · inspected 2026-07-20 · high confidence
Requires clarification when the target, outcome, scope, or success criteria are unclear.SKILL.md:lines 43-80 · inspected 2026-07-20 · high confidence
Still comparing?