Skip to content
Skills Library
Programming · Building & coding

refactor

/refactor

Code improvement workflow with polish and overhaul tracks. Triggers: 'refactor', 'clean up', 'restructure', 'reorganize', or /refactor. Phases: explore, brief, implement, validate. Existing code only — Do NOT use for bug fixes (/debug) or new features (/ideate).

Repository
exarchos ↗
Licence
Apache-2.0
Last repository change
24.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
6
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 24.10.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What it does and when to use it

Code improvement workflow with polish and overhaul tracks. Triggers: 'refactor', 'clean up', 'restructure', 'reorganize', or /refactor. Phases: explore, brief, implement, validate. Existing code only — Do NOT use for bug fixes (/debug) or new features (/ideate).

Triggers from the description

refactorclean uprestructurereorganize
Original author description

Code improvement workflow with polish and overhaul tracks. Triggers: 'refactor', 'clean up', 'restructure', 'reorganize', or /refactor. Phases: explore, brief, implement, validate. Existing code only — Do NOT use for bug fixes (/debug) or new features (/ideate).

How it works

  1. Overview
  2. Triggers
  3. Workflow Overview
  4. Command Interface
  5. Track Comparison
  6. When to use refactor vs debug
  7. Characterization Testing (Both Tracks)
  8. Design-time Constraints (Both Tracks)
  9. Polish Track
  10. Overhaul Track
Original SKILL.md headings · 204 lines, 1,428 words
SKILL.md preview
---
name: refactor
description: "Code improvement workflow with polish and overhaul tracks. Triggers: 'refactor', 'clean up', 'restructure', 'reorganize', or /refactor. Phases: explore, brief, implement, validate. Existing code only — Do NOT use for bug fixes (/debug) or new features (/ideate)."
metadata:
  author: exarchos
  version: 1.0.0
  mcp-server: exarchos
  category: workflow
  phase-affinity:
    - explore
    - brief
    - polish-implement
    - polish-validate
    - polish-update-docs
    - overhaul-plan
    - overhaul-plan-review
    - overhaul-delegate
    - overhaul-review
    - overhaul-update-docs
    - synthesize
---

# Refactor Workflow Skill

## Overview

Two-track workflow for improving existing code. Polish track for small, contained refactors; overhaul track for architectural changes and migrations. Both tracks emphasize exploration before commitment and mandatory documentation updates.

## Triggers

Activate this skill when:
- User runs `refactor` command
- User wants to improve existing code structure
- User mentions "refactor", "restructure", "clean up", "migrate"
- User asks to "move", "extract", "rename", or "reorganize" code

**Disambiguation:** If the user says "fix" or "clean up" — use `refactor` when the code *works* but needs structural improvement. Use `debug` when something is *broken* (error, crash, wrong behavior).

## Workflow Overview

```
                              refactor
                                  |
                            +-----+-----+
                            |  Explore  |
                            +-----+-----+
                                  |
                   +--------------+--------------+
                   |                             |
              --polish                       (default)
                   |                             |
                   v                             v
          +--------------+              +--------------+
          |    Polish    |              |   Overhaul   |
          |    Track     |              |    Track     |
          +--------------+              +--------------+
```

## Command Interface

### Start Refactor Workflow

```bash
# Default: overhaul track
refactor "Description of what needs refactoring"

# Fast path: polish track
refactor --polish "Small contained refactor description"

# Explore first, then decide track
refactor --explore "Unsure of scope, explore first"
```

### Mid-Workflow Commands

```bash
# Switch from polish to overhaul (during explore/brief)
refactor --switch-overhaul

# Resume after context compaction
rehydrate
```

## Track Comparison

| Aspect | Polish | Overhaul |
|--------|--------|----------|
| Scope | <=5 files, single concern | No limit |
| Worktree | No (direct) | Yes (isolated) |
| Delegation | No | Yes (full workflow) |
| Documentation | Mandatory update phase | Mandatory update phase |
| Human Checkpoints | 0 | 1 (merge) |

## When to use refactor vs debug

| Signal | Use refactor | Use debug |
|--------|--------------|-----------|
| Code works but is messy/complex | Yes | No |
| Something is broken or wrong | No | Yes |
| "This should be reorganized" | Yes | No |
| Users report a bug or regression | No | Yes |
| Performance degradation | Switch to refactor if structural | Start with debug (investigate) |
| SOLID violations in working code | Yes | No |
| Error in production logs | No | Yes |

**Rule of thumb:** if there is _dissatisfaction_ with working code (hard to read, violates SOLID, duplicated logic), use `refactor`. If there is a _symptom_ (something that should work but doesn't), use `debug`.

## Characterization Testing (Both Tracks)

Before modifying any existing code behavior, capture current behavior as characterization tests. This is a mandatory pre-step for both tracks:

1. **Before changes:** Write tests that document what the code **currently does** (not what it should do). Exercise the code through the most appropriate observable seam (API, CLI, integration boundary, or function) with representative inputs and assert on actual outputs/effects.
2. **During changes:** Any characterization test failure means behavior changed. Evaluate: intentional or accidental?
3. **After changes:** Document which characterization test failures were expected. Remaining characterization tests become regression tests.

This aligns with Michael Feathers' approach in *Working Effectively with Legacy Code* — understand behavior before changing it.

### Oracle-integrity gate (Both Tracks)

The characterization tests are your **oracle** — the safety net that tells you whether behavior changed. That net only holds if you do not edit the oracle mid-refactor. Before completing the refactor, inspect the test diff:

```bash
git diff -- tests/
```

The test oracle may be **ADDED to** during a refactor (new characterization tests, new cases), but an existing characterization assertion must **never be silently modified**. A silently-edited oracle invalidates the safety net: the test now passes against the *changed* behavior, hiding the very regression characterization was meant to catch. If a characterization assertion genuinely must change, that is a behavior change — call it out explicitly and justify it, do not fold it into the refactor diff unremarked.

## Design-time Constraints (Both Tracks)

At the **brief phase**, *before* committing to an approach, surface a **Constraints** section anchored to the architectural invariants relevant to the refactor. This is the refactor design-time equivalent of `/ideate`'s Phase 0 and uses the **same single shared source of truth** for the selection rules: see `@skills/ideate/references/constraint-anchoring.md`. Load `.exarchos/invariants.md` (`cost-of-load: always-load` entries) and emit the Constraints section. **Catalog-registration gated:** when no catalog is registered under `.exarchos.yml: invariants.catalogs`, surface no Constraints section and proceed directly. See `@skills/refactor/references/brief-template.md` for the brief-phase placement.

## Polish Track

Fast path for small, contained refactors (<=5 files, single concern). Orchestrator may write code directly (exception to orchestrator constraints). No worktree, no delegation.

HSM phases: `explore` → `brief` → `polish-implement` → `polish-validate` → `polish-update-docs` → `completed`

For detailed phase instructions, state management, and auto-chain behavior, see `@skills/refactor/references/polish-track.md`.

## Overhaul Track

Rigorous path for architectural changes, migrations, and multi-file restructuring. Uses full delegation model with worktree isolation.

HSM phases: `explore` → `brief` → `overhaul-plan` → `overhaul-plan-review` → `overhaul-delegate` → `overhaul-review` → `overhaul-update-docs` → `synthesize` → `completed`

For detailed phase instructions, skill invocations, and auto-chain behavior, see `@skills/refactor/references/overhaul-track.md`.

## State Management

Initialize refactor workflow:
```
action: "init", featureId: "refactor-<slug>", workflowType: "refactor"
```

Use `describe` to discover the full state schema at runtime: `exarchos_workflow({ action: "describe", actions: ["init"] })`.

### Phase Transitions and Guards

> **Sequential traversal required.** Every phase MUST be traversed in order — you cannot skip phases, even if you have all the data for a later phase ready. For example, `explore` must transition to `brief` before `overhaul-plan` — attempting `explore` → `overhaul-plan` directly will be rejected by the HSM. From `brief` you must go to `polish-implement` or `overhaul-plan`, not directly to `completed`. Each transition requires its guard to be satisfied via `updates` sent alongside the `phase` parameter in a single `set` call. See `@skills/refactor/references/polish-track.md` or `@skills/refactor/references/overhaul-track.md` for the exact tool call at each step.

Every phase transition has a guard that must be satisfied. Before transitioning, consult `@skills/checkpoint/references/phase-transitions.md` for the exact prerequisite for each guard.

The pattern for every transition: send the guard prerequisite in `updates` and the target in `phase` in a single `set` call.

### Schema Discovery

Use `exarchos_workflow({ action: "describe", actions: ["update", "init"] })` for
parameter schemas and `exarchos_workflow({ action: "describe", playbook: "refactor" })`
for phase transitions, guards, and playbook guidance.

### Decision Runbooks

For track-selection criteria at the explore phase, query the decision runbook:
`exarchos_orchestrate({ action: "runbook", id: "scope-decision" })`

This runbook provides structured criteria for choosing between polish and overhaul tracks based on file count, structural impact, and PR scope.

## Track Switching

If scope expands beyond polish limits during explore or brief phase, use `mcp__exarchos__exarchos_workflow` with `action: "update"` to set `track` to "overhaul" and update `explore.scopeAssessment.recommendedTrack`.

**Scope thresholds:** If >5 files affected OR changes cross module boundaries -> recommend overhaul track.

**Indicators to switch:**
- More than 5 files affected
- Multiple concerns identified
- Cross-module changes needed
- Test coverage gaps require new tests

Output: "Scope expanded beyond polish limits. Switching to overhaul track."

## Integration Points

**CRITICAL:** All skill invocations MUST use explicit `Skill()` tool calls:

| Skill | Invocation | Usage |
|-------|------------|-------|
| `plan` | `[Invoke the exarchos:plan skill with args: --refactor <state-file>]` | Task extraction from brief |
| `delegate` | `[Invoke the exarchos:delegate skill with args: <state-file>]` | Subagent dispatch for TDD |
| `review` | `[Invoke the exarchos:review skill with args: <state-file>]` | Quality review |
| `synthesize` | `[Invoke the exarchos:synthesize skill with args: <feature>]` | PR creation |

## Anti-Patterns

| Don't | Do Instead |
|-------|------------|
| Skip exploration | Always assess scope first (see `references/explore-checklist.md`) |
| Use polish for large changes | Switch to overhaul when scope expands |
| Skip doc updates | Documentation is mandatory (see `references/doc-update-checklist.md`) |
| Add features during refactor | Scope creep - stick to brief goals |
| Skip tests because "just moving code" | Refactors need test verification |
| Create design document for polish | Use brief in state file instead |
| Work in main for overhaul | Use worktree isolation |

## Exarchos Integration

When Exarchos MCP tools are available, emit events throughout the refactor workflow:

1. **At workflow start (explore):** `mcp__exarchos__exarchos_event` with `action: "append"` → `workflow.started` with workflowType "refactor"
2. **On track selection:** Auto-emitted by `exarchos_workflow` `set` when `phase` is provided — emits `workflow.transition` with selected track (polish/overhaul)
3. **On each phase transition:** Auto-emitted by `exarchos_workflow` `set` when `phase` is provided — emits `workflow.transition` with from/to/trigger/featureId
4. **Overhaul track stacking:** Handled by `delegate` (subagents use `git commit` + `git push` per implementer prompt)
5. **Polish track commit:** Single `git commit -m "refactor: <description>"` + `git push` — no multi-branch stacking needed
6. **On complete:** Auto-emitted by `exarchos_workflow` `set` when transitioning to terminal state — emits `workflow.transition` to "completed"
Files in this package (6)
  • SKILL.md 11 KB
  • references/brief-template.md 4 KB
  • references/doc-update-checklist.md 3 KB
  • references/explore-checklist.md 2 KB
  • references/overhaul-track.md 7 KB
  • references/polish-track.md 5 KB

Versions

RepositoryScoreLicence
This page · lvlup-sw/exarchos100Apache-2.0
froglet-studio/cosmic-shore100MIT
mgiovani/cc-arsenal100MIT
jongio/skills100MIT
boshu2/agentops100Apache-2.0
y-miyazaki/terraform-aws-base100Apache-2.0
mickeyyaya/evolve-loop100Apache-2.0
j0hanz/j0hanz-marketplace100MIT
kriscard/kriscard-claude-plugins100MIT

Identical content appears in 2 other GitHub folders. We show one copy with the highest score.

Database snapshot: 13.11.2026 · package version 7c9bae7e6e76 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100refactor-team
    Team-based code refactoring with analysis and user-approved changes. Use when: "refactor this code", "リファクタリング", "コード整理", "refactor team", "リファクタチーム".
  • 90implement
    Use this skill when writing code, building features, fixing bugs, refactoring, or performing multi-step implementation work. Activates on mentions of implement, build this, code this, start coding, fix this bug, refactor, make changes, develop this feature, coding task, write…
  • 100tdd
    Drive a portable red-green-refactor workflow for features, bug fixes, and changes to existing behavior. Use when implementing behavior with test-driven development, reproducing a bug as a failing test before fixing it, changing the semantics of existing code, or when the user…
  • 100codebase-cleanup-refactor-clean
    Refactor code with clean code principles, SOLID design patterns, and modern software engineering practices to improve quality, maintainability, and performance. Use when cleaning up or restructuring existing code.
  • 100refactor-plan
    Code-level refactoring analysis and roadmap. Detects duplication and complexity. Triggers on: refactor, tech debt, simplify, clean up, modularize, DRY, restructure.
  • 100workflow-refactor
    Scoped, behavior-preserving refactor: map dependencies, change structure, run affected tests. Use when "refactor this", "reduce duplication", or "improve readability". Dead code → plan-dead-code. Repo-wide smells → audit-code-quality.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.