Skip to content
cccskills
FreePublished on GitHub

refine-skill

Analyze and refine existing Claude Code Skills for conciseness and effectiveness. Use when asked to "review", "refine", "improve", "optimize" Skills, or when user mentions "skill quality", "skill refinement", "too verbose", "skill doesn't trigger", "description quality", or wants to make a Skill more concise. Do NOT use for creating new skills from scratch (use skill-creator instead).

The English summary is being prepared. Showing the original description.

See how to install

Files included (2)

  • SKILL.md4.9 KB
  • CHECKLIST.md5.0 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Skill Refinement

既存のClaude Code Skillを分析し、より簡潔で効果的なものに洗練する。

Core Principle

品質 = AIの消費しやすさ。Claude既知の情報は削除。

Workflow

  1. 読込: 全ファイル読み込み、行数・文字数測定
  2. 分析: CHECKLIST.mdと照合、問題特定
  3. 修正: 削除・追加・書き換え適用
  4. 検証: 行数減少、機能維持、トリガー明確化確認

Actions

  • 削除: 「〜とは」説明、背景説明、自明なステップ、重複
  • 追加: 具体例、トリガー、失敗対処
  • 書き換え: 曖昧→具体的、複雑→シンプル

Patterns

PatternAction
Description曖昧→具体的、トリガー追加
説明教育的散文→実行可能コード
ワークフロー10+ステップ→3-5ステップ
Progressive500行超→分割+参照

詳細例はCHECKLIST.md参照。

Quality Targets

  • Description: what + when + triggers (≤1024文字)
  • Lines: ≤500
  • References: 1階層
  • Examples: 具体的・実行可能

詳細: CHECKLIST.md

Description Formula

description は [What it does] + [When to use it] + [Key capabilities] の公式に従う。

# Good — 具体的な what/when/triggers
description: Extract text from PDFs, fill forms, merge documents. Use when working with PDF files or when user mentions "PDF", "forms", "document extraction".

# Bad — 曖昧、トリガーなし
description: Helps with documents.

# Bad — 技術的すぎ、ユーザー視点なし
description: Implements the Project entity model with hierarchical relationships.

negative trigger で over-triggering を防ぐ:

description: ...Use for statistical modeling, regression. Do NOT use for simple data exploration (use data-viz skill instead).

Common Issues

SymptomCauseFix
スキルがロードされないdescription が曖昧/トリガー不足trigger phrases 追加、ユーザーが言いそうなフレーズを含める
無関係なクエリでロードされるスコープが広すぎるnegative trigger 追加、対象を限定
指示が無視されるinstructions が冗長/埋もれている/曖昧重要指示をトップに、箇条書き化、詳細は references/ へ
出力品質が低いmodel laziness## Performance Notes で明示的に品質要求
過剰な説明Claude既知の情報を記載「〜とは」説明・背景・自明なステップを削除
深い参照チェーンA→B→C の多段参照1階層にフラット化

Iteration Signals

  • Undertriggering: ロードされない、手動有効化が必要、使い方の質問が来る → description にキーワード・ファイル種別を追加
  • Overtriggering: 無関係クエリでロード、無効化される → negative trigger 追加、スコープ限定
  • Execution issues: 一貫性なし、ユーザー修正必要 → instructions 改善、error handling 追加

Examples

Example 1: 冗長な description の改善

Before:

description: Creates sophisticated multi-page documentation systems.

問題: What はあるが When/Triggers がない。「sophisticated」は無意味。

After:

description: Generate multi-page documentation from source code. Use when user says "generate docs", "API documentation", or uploads code files for documentation.

Example 2: Instructions not followed の診断

Symptom: スキルはロードされるが指示通りに動かない。

診断チェック:

  1. Verbose? → 500行超なら references/ に分割
  2. Buried? → 重要指示が下部にある → ## Critical でトップに移動
  3. Ambiguous? → 「適切に処理」→ 具体的な検証項目リストに置換
  4. Model laziness? → ## Performance Notes セクションで品質明示

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Migrate existing ADRs to YAML frontmatter format. Scans docs/decisions/ for ADRs without frontmatter, infers status from body text, shows dry-run for approval. Trigger: /adr-migrate

The English summary is being prepared. Showing the original description.

berlysia/dotfiles9Updated Oct 21, 2026

Start or resume an ADR session. Without number: show status overview and recommend next actionable ADR. With number: provide phase-specific guidance. Trigger: /adr-session, /adr-session NNN, /adr-session new

The English summary is being prepared. Showing the original description.

berlysia/dotfiles9Updated Oct 21, 2026

Analyze a project's harness maturity and guide staged adoption of Harness Engineering best practices (quality loops, safety gates, linters, CLAUDE.md optimization).

The English summary is being prepared. Showing the original description.

berlysia/dotfiles9Updated Oct 21, 2026

bugfix

Free

Standardized bug fix workflow using TDD (Red-Green-Refactor). Use when fixing bugs, resolving issues, or when the user invokes /bugfix. Ensures root cause analysis, failing test first, then minimal fix.

The English summary is being prepared. Showing the original description.

berlysia/dotfiles9Updated Oct 21, 2026

Create detailed specifications by iteratively clarifying unclear points for Plan mode. Use when: After completing a plan when detailed requirements need clarification before implementation.

The English summary is being prepared. Showing the original description.

berlysia/dotfiles9Updated Oct 21, 2026

Policy-enforcing installer that deploys hardened Claude Code Action workflows (`@claude` manual mention + dependency-bot auto-fix) into a personal GitHub repository the user administers. Use when asked to introduce Claude Code Action, apply `@claude` mention support, or set up automated CI auto-fix for dependency bot PRs, using the security-hardened templates shipped with this skill.

The English summary is being prepared. Showing the original description.

berlysia/dotfiles9Updated Oct 21, 2026

See all skills by berlysia

Report a problem with this skill