skill-creator
Claude Codeスキル（SKILL.md）の設計・作成・バリデーション・レビュー。 Anthropic公式ガイド(2026-03)準拠。新規スキル作成、既存スキルの改善、 description品質チェック、トリガーテスト設計に使用。 「スキル作って」「スキル設計」「SKILL.md作成」「スキルレビュー」で起動。 Do NOT use for: スキルの実行・呼び出し（それは各スキル自体が行う）。
90
88%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a well-crafted skill description that clearly defines its scope (SKILL.md design, creation, validation, and review), provides explicit trigger terms in Japanese, and includes a helpful exclusion clause. It follows best practices by listing concrete actions, specifying when to use it, and distinguishing itself from related skills. The description uses appropriate third-person voice and is concise yet comprehensive.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 設計(design), 作成(creation), バリデーション(validation), レビュー(review), plus elaborates with 新規スキル作成, 既存スキルの改善, description品質チェック, トリガーテスト設計. These are concrete, actionable capabilities. | 3 / 3 |
Completeness | Clearly answers 'what' (design, create, validate, review SKILL.md files per Anthropic guidelines) and 'when' (explicit trigger phrases provided, plus a 'Do NOT use for' exclusion clause). Both dimensions are explicitly addressed with clear triggers. | 3 / 3 |
Trigger Term Quality | Includes natural trigger terms users would say in Japanese: 「スキル作って」「スキル設計」「SKILL.md作成」「スキルレビュー」. Also includes key terms like 'SKILL.md', 'スキル', 'description品質チェック' that cover common variations of how users would request this skill. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche: specifically about creating and reviewing SKILL.md files for Claude Code, not about executing skills or general coding. The 'Do NOT use for' clause further reduces conflict risk by explicitly excluding skill execution. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a comprehensive, well-structured meta-skill for creating Claude Code skills. Its greatest strengths are actionability (concrete examples, checklists, executable code) and workflow clarity (clear 12-step creation process with validation). Its main weakness is that it doesn't practice the progressive disclosure it preaches—the content is dense and monolithic at ~300 lines, with detailed reference material (frontmatter fields, design patterns, anti-patterns) that could be split into supporting files. Some verbosity could be trimmed where Claude's existing knowledge is assumed.
Suggestions
Split the detailed Frontmatter Reference, 5 Design Patterns, and Anti-Patterns table into references/ files (e.g., references/frontmatter.md, references/patterns.md) to practice the progressive disclosure the skill teaches and reduce SKILL.md length.
Remove explanations of concepts Claude already knows (e.g., what kebab-case means, basic YAML field descriptions like 'license: MIT # OSSスキル用。MIT, Apache-2.0 等') to improve conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is comprehensive but includes some content Claude already knows (e.g., explaining what kebab-case is, basic YAML syntax). The frontmatter reference section is thorough but could be more concise—many fields have obvious meanings. The anti-patterns table and some explanatory text add bulk. However, much of the domain-specific knowledge (description design, triggering patterns, progressive disclosure layers) genuinely adds value. | 2 / 3 |
Actionability | Highly actionable with concrete examples throughout: the description design section has good/bad YAML examples, the 7-item checklist is specific and usable, the creation workflow is a numbered step-by-step process, validation script examples are executable bash, and the frontmatter reference provides copy-paste-ready YAML with all fields documented. The triggering test examples are concrete and testable. | 3 / 3 |
Workflow Clarity | The 12-step Creation Workflow is clearly sequenced with logical dependencies. The testing strategy covers 3 distinct areas with explicit examples. The description design includes a debugging feedback loop ('ask Claude when would you use this skill'). Validation is explicitly called out as step 10 with a concrete script example. The progressive disclosure layers (L1/L2/L3) clearly define when content is loaded. | 3 / 3 |
Progressive Disclosure | The skill itself teaches progressive disclosure (L1/L2/L3 layers, references/ directory structure) but doesn't practice it well—it's a ~300-line monolithic file with no bundle files to offload detailed content. The frontmatter reference, all 5 design patterns, and the anti-patterns table could be split into references/ files. The file structure section mentions references/ and scripts/ directories but none are provided. | 2 / 3 |
Total | 10 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- yohey-w/multi-agent-shogun
- Commit
4ee1377
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.