From evolutionary-naming
Walks code through a 7-step naming process to improve identifiers with poor names. Two modes: audit (lists naming opportunities broadly) and improve (walks one identifier through the process).
How this skill is triggered — by the user, by Claude, or both
Slash command
/evolutionary-naming:evolutionary-namingThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Naming is a process, not a single step.** Names evolve through 7 progressive steps grouped into 3 phases. Match the depth of work to the user's actual story — don't run the full pipeline every time.
Naming is a process, not a single step. Names evolve through 7 progressive steps grouped into 3 phases. Match the depth of work to the user's actual story — don't run the full pipeline every time.
This skill is advisory. It reads code and proposes renames, transitions, and the commit message you'd write for each — it does not edit files or run git. The user applies the changes. Phrase every transition as a proposal, and suggest the commit message they would use, rather than claiming the rename is done.
Based on Arlo Belshee's "Naming as a Process" (CC BY 3.0). Updated edition: digdeeproots.com.
| Phase | Steps | Nature |
|---|---|---|
| Phase 1: Insight → Name | Missing → Nonsense → Honest → Honest and Complete | Universal. Pure naming. No structural change. Safe to walk continuously. |
| Phase 2: Name → Structure | Honest and Complete → Does the Right Thing | Codebase-specific. Requires structural refactoring. Ask permission. |
| Phase 3: Combine for Design | Does the Right Thing → Intent → Domain Abstraction | Requires reading call sites and domain context. Ask permission. |
This skill operates in two modes. Pick the one that matches the user's request.
| User signal | Mode | File |
|---|---|---|
Specific identifier named ("d を改善", "DocumentManager をどうにか", "this method") | improve-mode | improve-mode.md |
| One clear target inside a code block + "リファクタリング" / "改善" | improve-mode | improve-mode.md |
| "全体の命名を見直したい", "改善余地ある?", "命名レビュー", "scan", "audit", "リストだけほしい" | audit-mode | audit-mode.md |
| Audit followup: "X だけ直して" after audit output | improve-mode (target = X) | improve-mode.md |
| Ambiguous (large code paste, no specific target, no "全体" language) | Ask the user: "全体を監査しますか、特定の識別子を改善しますか?" |
Mode = audit: Read audit-mode.md. Produce a structured table of all naming opportunities. Do not execute changes.
Mode = improve: Read improve-mode.md. Walk one identifier through phase transitions. Pause at phase boundaries.
Both modes share diagnostic criteria, transition mechanics, and red flags from reference.md.
How far to walk depends on the user's depth signal. The canonical signal → stop-at table lives in reference.md ("Depth Signals — How Far to Walk"); both modes read from it.
"You take the minimum steps to get the name to meet your need for the current story." — leaving names incomplete for colleagues to extend is a feature, not a flaw.
SKILL.md — this file. Router and overview.reference.md — shared 7-step diagnosis table, transitions, red flags, applesauce/probably_/AndStuff guidance.audit-mode.md — exhaustive scan workflow + output template.improve-mode.md — interactive single-target workflow + pause protocol.Read the mode file you selected, plus reference.md for diagnosis details.
npx claudepluginhub kawasima/evolutionary-naming --plugin evolutionary-namingSuggests improved names for variables, functions, classes, files, DB tables, and API endpoints using language-specific conventions like camelCase or snake_case.
Choosing meaningful, pronounceable names that reveal intent for functions, variables, classes, and modules.
Enforces precise naming for variables, functions, classes, files, and identifiers. Bans vague names like data/temp/result/info/handle/manager; promotes semantic accuracy and domain-specific terms via specificity ladder.