create-ai-project 1.11.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (150) hide show
  1. package/.claude/agents/acceptance-test-generator.md +316 -0
  2. package/.claude/agents/code-reviewer.md +193 -0
  3. package/.claude/agents/document-reviewer.md +182 -0
  4. package/.claude/agents/prd-creator.md +186 -0
  5. package/.claude/agents/quality-fixer.md +295 -0
  6. package/.claude/agents/requirement-analyzer.md +161 -0
  7. package/.claude/agents/rule-advisor.md +194 -0
  8. package/.claude/agents/task-decomposer.md +291 -0
  9. package/.claude/agents/task-executor.md +270 -0
  10. package/.claude/agents/technical-designer.md +343 -0
  11. package/.claude/agents/work-planner.md +181 -0
  12. package/.claude/agents-en/acceptance-test-generator.md +256 -0
  13. package/.claude/agents-en/code-reviewer.md +195 -0
  14. package/.claude/agents-en/design-sync.md +225 -0
  15. package/.claude/agents-en/document-reviewer.md +190 -0
  16. package/.claude/agents-en/integration-test-reviewer.md +195 -0
  17. package/.claude/agents-en/prd-creator.md +196 -0
  18. package/.claude/agents-en/quality-fixer-frontend.md +334 -0
  19. package/.claude/agents-en/quality-fixer.md +291 -0
  20. package/.claude/agents-en/requirement-analyzer.md +165 -0
  21. package/.claude/agents-en/rule-advisor.md +194 -0
  22. package/.claude/agents-en/task-decomposer.md +291 -0
  23. package/.claude/agents-en/task-executor-frontend.md +276 -0
  24. package/.claude/agents-en/task-executor.md +272 -0
  25. package/.claude/agents-en/technical-designer-frontend.md +441 -0
  26. package/.claude/agents-en/technical-designer.md +371 -0
  27. package/.claude/agents-en/work-planner.md +216 -0
  28. package/.claude/agents-ja/acceptance-test-generator.md +256 -0
  29. package/.claude/agents-ja/code-reviewer.md +195 -0
  30. package/.claude/agents-ja/design-sync.md +225 -0
  31. package/.claude/agents-ja/document-reviewer.md +192 -0
  32. package/.claude/agents-ja/integration-test-reviewer.md +195 -0
  33. package/.claude/agents-ja/prd-creator.md +194 -0
  34. package/.claude/agents-ja/quality-fixer-frontend.md +335 -0
  35. package/.claude/agents-ja/quality-fixer.md +292 -0
  36. package/.claude/agents-ja/requirement-analyzer.md +164 -0
  37. package/.claude/agents-ja/rule-advisor.md +194 -0
  38. package/.claude/agents-ja/task-decomposer.md +291 -0
  39. package/.claude/agents-ja/task-executor-frontend.md +276 -0
  40. package/.claude/agents-ja/task-executor.md +272 -0
  41. package/.claude/agents-ja/technical-designer-frontend.md +442 -0
  42. package/.claude/agents-ja/technical-designer.md +370 -0
  43. package/.claude/agents-ja/work-planner.md +213 -0
  44. package/.claude/commands/build.md +78 -0
  45. package/.claude/commands/design.md +27 -0
  46. package/.claude/commands/implement.md +79 -0
  47. package/.claude/commands/plan.md +43 -0
  48. package/.claude/commands/project-inject.md +76 -0
  49. package/.claude/commands/refine-rule.md +206 -0
  50. package/.claude/commands/review.md +78 -0
  51. package/.claude/commands/sync-rules.md +116 -0
  52. package/.claude/commands/task.md +13 -0
  53. package/.claude/commands-en/build.md +77 -0
  54. package/.claude/commands-en/design.md +39 -0
  55. package/.claude/commands-en/front-build.md +103 -0
  56. package/.claude/commands-en/front-design.md +42 -0
  57. package/.claude/commands-en/front-plan.md +40 -0
  58. package/.claude/commands-en/implement.md +75 -0
  59. package/.claude/commands-en/plan.md +45 -0
  60. package/.claude/commands-en/project-inject.md +76 -0
  61. package/.claude/commands-en/refine-rule.md +208 -0
  62. package/.claude/commands-en/review.md +78 -0
  63. package/.claude/commands-en/sync-rules.md +116 -0
  64. package/.claude/commands-en/task.md +13 -0
  65. package/.claude/commands-ja/build.md +75 -0
  66. package/.claude/commands-ja/design.md +37 -0
  67. package/.claude/commands-ja/front-build.md +103 -0
  68. package/.claude/commands-ja/front-design.md +42 -0
  69. package/.claude/commands-ja/front-plan.md +40 -0
  70. package/.claude/commands-ja/implement.md +73 -0
  71. package/.claude/commands-ja/plan.md +43 -0
  72. package/.claude/commands-ja/project-inject.md +76 -0
  73. package/.claude/commands-ja/refine-rule.md +206 -0
  74. package/.claude/commands-ja/review.md +78 -0
  75. package/.claude/commands-ja/sync-rules.md +116 -0
  76. package/.claude/commands-ja/task.md +13 -0
  77. package/.claude/settings.local.json +74 -0
  78. package/.husky/pre-commit +1 -0
  79. package/.husky/pre-push +3 -0
  80. package/.madgerc +14 -0
  81. package/.tsprunerc +11 -0
  82. package/CLAUDE.en.md +102 -0
  83. package/CLAUDE.ja.md +102 -0
  84. package/CLAUDE.md +111 -0
  85. package/LICENSE +21 -0
  86. package/README.ja.md +233 -0
  87. package/README.md +243 -0
  88. package/bin/create-project.js +87 -0
  89. package/biome.json +51 -0
  90. package/docs/adr/template-en.md +64 -0
  91. package/docs/adr/template-ja.md +64 -0
  92. package/docs/design/template-en.md +281 -0
  93. package/docs/design/template-ja.md +285 -0
  94. package/docs/guides/en/quickstart.md +111 -0
  95. package/docs/guides/en/rule-editing-guide.md +266 -0
  96. package/docs/guides/en/sub-agents.md +343 -0
  97. package/docs/guides/en/use-cases.md +308 -0
  98. package/docs/guides/ja/quickstart.md +112 -0
  99. package/docs/guides/ja/rule-editing-guide.md +266 -0
  100. package/docs/guides/ja/sub-agents.md +343 -0
  101. package/docs/guides/ja/use-cases.md +290 -0
  102. package/docs/guides/sub-agents.md +306 -0
  103. package/docs/plans/20250123-integration-test-improvement.md +993 -0
  104. package/docs/plans/template-en.md +130 -0
  105. package/docs/plans/template-ja.md +130 -0
  106. package/docs/prd/template-en.md +109 -0
  107. package/docs/prd/template-ja.md +109 -0
  108. package/docs/rules/ai-development-guide.md +260 -0
  109. package/docs/rules/architecture/implementation-approach.md +136 -0
  110. package/docs/rules/documentation-criteria.md +180 -0
  111. package/docs/rules/project-context.md +38 -0
  112. package/docs/rules/rules-index.yaml +137 -0
  113. package/docs/rules/technical-spec.md +47 -0
  114. package/docs/rules/typescript-testing.md +188 -0
  115. package/docs/rules/typescript.md +166 -0
  116. package/docs/rules-en/architecture/implementation-approach.md +136 -0
  117. package/docs/rules-en/coding-standards.md +333 -0
  118. package/docs/rules-en/documentation-criteria.md +184 -0
  119. package/docs/rules-en/frontend/technical-spec.md +143 -0
  120. package/docs/rules-en/frontend/typescript-testing.md +124 -0
  121. package/docs/rules-en/frontend/typescript.md +131 -0
  122. package/docs/rules-en/integration-e2e-testing.md +149 -0
  123. package/docs/rules-en/project-context.md +38 -0
  124. package/docs/rules-en/rules-index.yaml +211 -0
  125. package/docs/rules-en/technical-spec.md +86 -0
  126. package/docs/rules-en/typescript-testing.md +149 -0
  127. package/docs/rules-en/typescript.md +116 -0
  128. package/docs/rules-ja/architecture/implementation-approach.md +136 -0
  129. package/docs/rules-ja/coding-standards.md +333 -0
  130. package/docs/rules-ja/documentation-criteria.md +180 -0
  131. package/docs/rules-ja/frontend/technical-spec.md +143 -0
  132. package/docs/rules-ja/frontend/typescript-testing.md +124 -0
  133. package/docs/rules-ja/frontend/typescript.md +131 -0
  134. package/docs/rules-ja/integration-e2e-testing.md +149 -0
  135. package/docs/rules-ja/project-context.md +38 -0
  136. package/docs/rules-ja/rules-index.yaml +196 -0
  137. package/docs/rules-ja/technical-spec.md +86 -0
  138. package/docs/rules-ja/typescript-testing.md +149 -0
  139. package/docs/rules-ja/typescript.md +116 -0
  140. package/package.json +98 -0
  141. package/scripts/check-unused-exports.js +69 -0
  142. package/scripts/cleanup-test-processes.sh +32 -0
  143. package/scripts/post-setup.js +110 -0
  144. package/scripts/set-language.js +310 -0
  145. package/scripts/setup-project.js +199 -0
  146. package/scripts/show-coverage.js +74 -0
  147. package/src/index.ts +11 -0
  148. package/templates/.gitignore.template +52 -0
  149. package/tsconfig.json +50 -0
  150. package/vitest.config.mjs +47 -0
@@ -0,0 +1,74 @@
1
+ {
2
+ "$schema": "https://json.schemastore.org/claude-code-settings.json",
3
+ "permissions": {
4
+ "allow": [
5
+ "Bash(npm run check:unused:*)",
6
+ "Read(/Users/kagawashinsuke/Documents/tacoms-ai-bot/scripts/**)",
7
+ "WebSearch",
8
+ "Bash(npx ts-prune:*)",
9
+ "Bash(git checkout:*)",
10
+ "Bash(git pull:*)",
11
+ "Bash(chmod:*)",
12
+ "Bash(git add:*)",
13
+ "Bash(git commit:*)",
14
+ "Bash(gh api:*)",
15
+ "Bash(git branch:*)",
16
+ "WebFetch(domain:github.com)",
17
+ "Bash(npm install:*)",
18
+ "Bash(git reset:*)",
19
+ "WebFetch(domain:qiita.com)",
20
+ "Bash(cat:*)",
21
+ "Bash(git push:*)",
22
+ "WebFetch(domain:mermaid.js.org)",
23
+ "Bash(gh pr view:*)",
24
+ "Bash(gh pr diff:*)",
25
+ "Read(//tmp/**)",
26
+ "Bash(gh repo clone:*)",
27
+ "Bash(git stash:*)",
28
+ "Bash(npm run lang:ja:*)",
29
+ "Bash(find:*)",
30
+ "Bash(awk:*)",
31
+ "Bash(grep:*)",
32
+ "Bash(xargs basename:*)",
33
+ "WebFetch(domain:dev.to)",
34
+ "Bash(gh pr checks:*)",
35
+ "WebFetch(domain:www.anthropic.com)",
36
+ "Read(//Users/kagawashinsuke/Documents/agentic-code/**)",
37
+ "Read(//Users/kagawashinsuke/Documents/claude-code-workflows/**)",
38
+ "Bash(git -C /Users/kagawashinsuke/Documents/claude-code-workflows diff agents/acceptance-test-generator.md)",
39
+ "Read(//Users/kagawashinsuke/Documents/project-qb/**)",
40
+ "Bash(git rm:*)",
41
+ "Bash(git ls-tree:*)",
42
+ "Bash(node bin/create-project.js:*)",
43
+ "Bash(node:*)",
44
+ "WebFetch(domain:martinfowler.com)",
45
+ "Bash(test:*)",
46
+ "WebFetch(domain:www.agileconnection.com)",
47
+ "Bash(git log:*)",
48
+ "Bash(git mv:*)",
49
+ "Bash(gh pr create:*)",
50
+ "Bash(xargs git branch -D)",
51
+ "Bash(ls:*)",
52
+ "Bash(git clone:*)",
53
+ "WebFetch(domain:support.atlassian.com)",
54
+ "WebFetch(domain:community.atlassian.com)",
55
+ "WebFetch(domain:www.velir.com)",
56
+ "WebFetch(domain:scottspence.com)",
57
+ "WebFetch(domain:arxiv.org)",
58
+ "WebFetch(domain:dl.acm.org)",
59
+ "Bash(gh issue create:*)",
60
+ "Bash(git fetch:*)",
61
+ "Bash(git diff:*)",
62
+ "Bash(fi)",
63
+ "Bash(done)",
64
+ "WebFetch(domain:www.npmjs.com)",
65
+ "Bash(npm pack:*)",
66
+ "Bash(npm view:*)",
67
+ "Bash(npx:*)",
68
+ "Bash(=20\" for compatibility\n\n🤖 Generated with [Claude Code](https://claude.com/claude-code)\n\nCo-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>\nEOF\n)\")"
69
+ ],
70
+ "deny": [],
71
+ "ask": [],
72
+ "defaultMode": "acceptEdits"
73
+ }
74
+ }
@@ -0,0 +1 @@
1
+ npx lint-staged
@@ -0,0 +1,3 @@
1
+ #!/bin/sh
2
+ npm run check
3
+ npm run build
package/.madgerc ADDED
@@ -0,0 +1,14 @@
1
+ {
2
+ "fileExtensions": ["ts"],
3
+ "excludeRegExp": [
4
+ "__tests__",
5
+ "\\.test\\.ts$",
6
+ "\\.spec\\.ts$"
7
+ ],
8
+ "detectiveOptions": {
9
+ "ts": {
10
+ "skipTypeImports": true
11
+ }
12
+ },
13
+ "tsConfig": "./tsconfig.json"
14
+ }
package/.tsprunerc ADDED
@@ -0,0 +1,11 @@
1
+ {
2
+ "ignore": [
3
+ "src/index.ts",
4
+ "**/__tests__/**",
5
+ "**/*.test.ts",
6
+ "**/*.spec.ts",
7
+ "**/test/**",
8
+ "**/vitest.config.ts",
9
+ "**/vitest.setup.ts"
10
+ ]
11
+ }
package/CLAUDE.en.md ADDED
@@ -0,0 +1,102 @@
1
+ # Claude Code Development Rules
2
+
3
+ Core rules for maximizing AI execution accuracy. All instructions must follow this file.
4
+
5
+ ## 🚨 Most Important Principle: Research OK, Implementation STOP
6
+
7
+ **User approval is mandatory before using any Edit/Write/MultiEdit tools**
8
+
9
+ Reason: To prevent implementations that differ from user intent and ensure correct direction
10
+
11
+ ## Mandatory Execution Process
12
+
13
+ ### Execution Flow (Required Steps)
14
+ 1. **Task decomposition with TodoWrite** → Cannot proceed to implementation without it
15
+ Reason: To structure tasks and enable progress tracking
16
+ 2. **Execute rule-advisor** → Understand task essence and select appropriate rules
17
+ Reason: To select appropriate rules and understand task essence
18
+ 3. **Use Edit/Write/MultiEdit** → User approval is mandatory
19
+ Reason: Practicing the most important principle (Research OK, Implementation STOP)
20
+ 4. **Execute implementation** → Cannot complete with quality check errors
21
+ Reason: To ensure high-quality code
22
+ 5. **Count file changes** → Auto-stop when threshold exceeded
23
+ Reason: To get confirmation before impact becomes too large
24
+
25
+ ### TodoWrite and Metacognition Integration
26
+ **Execution Rules**:
27
+ - When `pending → in_progress`: rule-advisor output is mandatory
28
+ - **After rule-advisor execution**: Always update TodoWrite in the following format
29
+ 1. Add firstActionGuidance as the first task in Todo
30
+ 2. Record taskEssence as the completion criteria for each task
31
+ 3. Record warningPatterns as checkpoints during execution
32
+ - Using Edit tools without TodoWrite: Stop as rule violation
33
+ - When updating task status: Recording implementation details is mandatory (no blanks)
34
+
35
+ ### Execution Prerequisites
36
+ 1. **rule-advisor agent execution (invoke using Task tool)** → JSON response must exist
37
+ 2. **TodoWrite task** → in_progress status must exist
38
+ 3. **User approval record** → Explicit approval before Edit/Write
39
+ 4. **Quality check results** → Cannot complete with errors > 0
40
+
41
+ ### Required Alternative Patterns
42
+ - **Instead of any type** → Use unknown type with type guards
43
+ Reason: To ensure type safety and prevent runtime errors
44
+ - **When using Edit** → Always precede with TodoWrite for task management
45
+ Reason: To enable progress tracking and quality assurance
46
+ - **When using Edit/Write/MultiEdit** → Obtain user approval before execution
47
+ Reason: To comply with the most important principle (Research OK, Implementation STOP)
48
+ - **When declaring completion** → Confirm zero quality check errors before declaration
49
+ Reason: To merge only complete code
50
+
51
+ ## Metacognitive Execution (Mandatory at Task Start)
52
+
53
+ ### Metacognition with Information from rule-advisor
54
+ 1. **Understand taskEssence (task essence)**
55
+ - Distinguish between surface work and fundamental purpose
56
+ - Judge "quick fix" vs "proper solution"
57
+
58
+ 2. **Confirm applicableRules (applicable rules)**
59
+ - Judge if selected rules are appropriate
60
+ - Load necessary sections
61
+
62
+ 3. **Recognize pastFailurePatterns (past failures)**
63
+ - Be careful not to repeat same failures
64
+ - Be conscious of suggested workarounds
65
+
66
+ 4. **Execute firstAction (initial action)**
67
+ - Start with recommended tools
68
+ - Proceed systematically
69
+
70
+ ## Claude Behavior Control (Failure Prevention)
71
+
72
+ ### Auto-stop Triggers (Must Stop)
73
+ - **Detecting 5+ file changes**: Stop immediately, report impact to user
74
+ Reason: Large changes require advance planning and review
75
+ - **Same error occurs 3 times**: Root cause analysis mandatory
76
+ Reason: Need root solution, not symptomatic treatment
77
+ - **Detecting excessive unknown type usage**: Reconsider type guard design
78
+ Reason: High possibility of type safety design issues
79
+ - **After editing 3 files**: Force TodoWrite update (cannot use next Edit tool without update)
80
+ Reason: Need to confirm progress and direction
81
+
82
+ ### Handling Error-fixing Impulse
83
+ 1. Error discovered → **Pause**
84
+ 2. Re-execute rule-advisor
85
+ 3. Root cause analysis (repeat "why" 5 times to identify true cause)
86
+ 4. Present action plan
87
+ 5. Fix after user approval
88
+
89
+ ### Escalation Criteria (User Confirmation Required)
90
+ - Architecture changes (adding new layers, changing responsibilities)
91
+ - Adding external dependencies (npm packages, external APIs)
92
+ - Breaking changes (changing existing APIs, data structures)
93
+ - Multiple implementation methods with unclear superiority
94
+
95
+ ### Preventing Rule Ignoring When Focused
96
+ **Measurable Forced Stop Criteria**:
97
+ - **2nd consecutive error fix**: Automatically trigger rule-advisor re-execution
98
+ - **5 Edit tool uses**: Force impact report creation
99
+ - **3 edits to same file**: Force stop for refactoring consideration
100
+
101
+ ### Temporary File Creation Rules
102
+ Use `tmp/` directory for work files. Delete upon completion.
package/CLAUDE.ja.md ADDED
@@ -0,0 +1,102 @@
1
+ # Claude Code 開発ルール
2
+
3
+ AI実行精度最大化のための中核ルール。全ての指示はこのファイルに従う。
4
+
5
+ ## 🚨 最重要原則:調査OK、実装STOP
6
+
7
+ **すべてのEdit/Write/MultiEditツール使用前にユーザー承認が必須**
8
+
9
+ 理由:ユーザーの意図と異なる実装を防ぎ、正しい方向性を確保するため
10
+
11
+ ## 必須実行プロセス
12
+
13
+ ### 実行フロー(必須手順)
14
+ 1. **TodoWriteでタスク分解** → なければ実装ステップに進めない
15
+ 理由:タスクを構造化し、進捗を追跡可能にするため
16
+ 2. **rule-advisor実行** → タスクの本質を理解し適切なルールを選択
17
+ 理由:適切なルールを選択し、タスクの本質を理解するため
18
+ 3. **Edit/Write/MultiEdit使用** → ユーザー承認が必須
19
+ 理由:最重要原則の実践(調査OK、実装STOP)
20
+ 4. **実装実行** → 品質チェックエラーは完了条件を満たさない
21
+ 理由:高品質なコードを保証するため
22
+ 5. **ファイル数カウント** → 閾値超過で自動停止
23
+ 理由:影響範囲が大きくなる前に確認を取るため
24
+
25
+ ### TodoWriteとメタ認知の統合
26
+ **実行ルール**:
27
+ - `pending → in_progress`時: rule-advisorの出力が必須
28
+ - **rule-advisor実行後**: 必ずTodoWriteを以下の形式で更新
29
+ 1. firstActionGuidanceをTodoの先頭タスクとして追加
30
+ 2. taskEssenceを各タスクの完了判断基準として記録
31
+ 3. warningPatternsを実行中の確認項目として記録
32
+ - TodoWriteなしでEditツール使用: ルール違反として停止
33
+ - 各タスクのステータス更新時: 実施内容の記録を必須化(空白不可)
34
+
35
+ ### 実行前提条件
36
+ 1. **rule-advisorエージェント実行(Taskツールで呼び出すこと)** → JSONレスポンスが存在すること
37
+ 2. **TodoWriteのタスク** → in_progressステータスが存在すること
38
+ 3. **ユーザー承認記録** → Edit/Write前に明示的承認があること
39
+ 4. **品質チェック結果** → エラー0でなければ完了不可
40
+
41
+ ### 必須の代替パターン
42
+ - **any型の代わり** → unknown型と型ガードを使用
43
+ 理由:型安全性を確保し、実行時エラーを防止するため
44
+ - **Edit使用時** → 必ずTodoWriteでタスク管理を先行
45
+ 理由:進捗追跡と品質保証を可能にするため
46
+ - **Edit/Write/MultiEdit使用時** → ユーザー承認を取得してから実行
47
+ 理由:最重要原則(調査OK、実装STOP)を遵守するため
48
+ - **完了宣言時** → 品質チェックエラー0を確認してから宣言
49
+ 理由:完全なコードのみをマージするため
50
+
51
+ ## メタ認知実行(タスク開始時必須)
52
+
53
+ ### rule-advisorから取得した情報でメタ認知
54
+ 1. **taskEssence(タスクの本質)を理解**
55
+ - 表面的な作業内容と根本目的の区別
56
+ - 「quick fix」vs「proper solution」の判定
57
+
58
+ 2. **applicableRules(適用ルール)を確認**
59
+ - 選択されたルールが適切か判断
60
+ - 必要なセクションを読み込む
61
+
62
+ 3. **pastFailurePatterns(過去の失敗)を認識**
63
+ - 同じ失敗を繰り返さないよう注意
64
+ - 提示された回避策を意識
65
+
66
+ 4. **firstAction(初動アクション)を実行**
67
+ - 推奨されたツールから開始
68
+ - 計画的に進める
69
+
70
+ ## Claude行動制御(失敗防止)
71
+
72
+ ### 自動停止トリガー(必ず停止)
73
+ - **5ファイル以上の変更検出**:即座停止、影響範囲をユーザーに報告
74
+ 理由:大規模変更は事前計画とレビューが必要なため
75
+ - **同じエラー3回発生**:根本原因分析必須
76
+ 理由:対症療法ではなく根本解決が必要なため
77
+ - **unknown型多用検出**:型ガード設計再検討
78
+ 理由:型安全性の設計に問題がある可能性が高いため
79
+ - **3ファイル編集完了**: TodoWriteの更新を強制(更新しないと次のEditツール使用不可)
80
+ 理由:進捗確認と方向性の再確認が必要なため
81
+
82
+ ### エラー修正衝動対処
83
+ 1. エラー発見 → **一時停止**
84
+ 2. rule-advisor再実行
85
+ 3. 根本原因分析(なぜ?を5回繰り返して真因を特定)
86
+ 4. 対処計画提示
87
+ 5. ユーザー承認後に修正
88
+
89
+ ### エスカレーション基準(ユーザー確認必須)
90
+ - アーキテクチャ変更(新レイヤー追加、責務変更)
91
+ - 外部依存追加(npmパッケージ、外部API)
92
+ - 破壊的変更(既存APIの変更、データ構造の変更)
93
+ - 複数の実装方法があり優劣が判断できない場合
94
+
95
+ ### 集中時のルール無視防止
96
+ **測定可能な強制停止基準**:
97
+ - **連続エラー修正2回目**: 自動的にrule-advisor再実行をトリガー
98
+ - **Editツール5回使用**: 影響範囲レポートの作成を強制
99
+ - **同一ファイル3回編集**: リファクタリング検討の強制停止
100
+
101
+ ### 一時ファイル作成ルール
102
+ 作業中ファイルは`tmp/`ディレクトリ使用。完了時削除。
package/CLAUDE.md ADDED
@@ -0,0 +1,111 @@
1
+ # Claude Code 開発ルール
2
+
3
+ AI実行精度最大化のための中核ルール。全ての指示はこのファイルに従う。
4
+
5
+ ## 🚨 最重要原則:調査OK、実装STOP
6
+
7
+ **すべてのEdit/Write/MultiEditツール使用前にユーザー承認が必須**
8
+
9
+ 理由:ユーザーの意図と異なる実装を防ぎ、正しい方向性を確保するため
10
+
11
+ ## 必須実行プロセス
12
+
13
+ ### 実行フロー(必須手順)
14
+ 1. **TodoWriteでタスク分解** → なければ実装ステップに進めない
15
+ 理由:タスクを構造化し、進捗を追跡可能にするため
16
+ 2. **rule-advisor実行** → タスクの本質を理解し適切なルールを選択
17
+ 理由:適切なルールを選択し、タスクの本質を理解するため
18
+ 3. **Edit/Write/MultiEdit使用** → ユーザー承認が必須
19
+ 理由:最重要原則の実践(調査OK、実装STOP)
20
+ 4. **実装実行** → 品質チェックエラーは完了条件を満たさない
21
+ 理由:高品質なコードを保証するため
22
+ 5. **ファイル数カウント** → 閾値超過で自動停止
23
+ 理由:影響範囲が大きくなる前に確認を取るため
24
+
25
+ ### TodoWriteとメタ認知の統合
26
+ **実行ルール**:
27
+ - `pending → in_progress`時: rule-advisorの出力が必須
28
+ - **rule-advisor実行後**: 必ずTodoWriteを更新(タスク内容・優先度・分解粒度の見直し)
29
+ - taskEssenceに基づいてタスク説明を具体化
30
+ - firstActionGuidanceを最初のタスクに反映
31
+ - warningPatternsを考慮してタスクを再構成
32
+ - TodoWriteなしでEditツール使用: ルール違反として停止
33
+ - 各タスクのステータス更新時: 実施内容の記録を必須化(空白不可)
34
+
35
+ ### 実行前提条件
36
+ 1. **rule-advisorの出力** → JSONレスポンスが存在すること
37
+ 2. **TodoWriteのタスク** → in_progressステータスが存在すること
38
+ 3. **ユーザー承認記録** → Edit/Write前に明示的承認があること
39
+ 4. **品質チェック結果** → エラー0でなければ完了不可
40
+
41
+ ### 禁止事項
42
+ - **any型使用** → 型チェックエラーの原因
43
+ 理由:型安全性を破壊し、実行時エラーを引き起こすため
44
+ - **TodoWriteなしのEdit使用** → ルール違反
45
+ 理由:タスク管理なしでは進捗追跡と品質保証ができないため
46
+ - **承認なしEdit/Write/MultiEdit** → 承認待ちに移行
47
+ 理由:最重要原則(調査OK、実装STOP)の違反を防ぐため
48
+ - **品質エラー残存での完了宣言** → 完了条件を満たさない
49
+ 理由:不完全なコードの混入を防ぐため
50
+
51
+ ## メタ認知実行(タスク開始時必須)
52
+
53
+ ### rule-advisorから取得した情報でメタ認知
54
+ 1. **taskEssence(タスクの本質)を理解**
55
+ - 表面的な作業内容と根本目的の区別
56
+ - 「quick fix」vs「proper solution」の判定
57
+
58
+ 2. **applicableRules(適用ルール)を確認**
59
+ - 選択されたルールが適切か判断
60
+ - 必要なセクションを読み込む
61
+
62
+ 3. **pastFailurePatterns(過去の失敗)を認識**
63
+ - 同じ失敗を繰り返さないよう注意
64
+ - 提示された回避策を意識
65
+
66
+ 4. **firstAction(初動アクション)を実行**
67
+ - 推奨されたツールから開始
68
+ - 計画的に進める
69
+
70
+ ## Claude行動制御(失敗防止)
71
+
72
+ ### 自動停止トリガー(必ず停止)
73
+ - **5ファイル以上の変更検出**:即座停止、影響範囲をユーザーに報告
74
+ 理由:大規模変更は事前計画とレビューが必要なため
75
+ - **同じエラー3回発生**:根本原因分析必須
76
+ 理由:対症療法ではなく根本解決が必要なため
77
+ - **unknown型多用検出**:型ガード設計再検討
78
+ 理由:型安全性の設計に問題がある可能性が高いため
79
+ - **3ファイル編集完了**: TodoWriteの更新を強制(更新しないと次のEditツール使用不可)
80
+ 理由:進捗確認と方向性の再確認が必要なため
81
+
82
+ ### エラー修正衝動対処
83
+ 1. エラー発見 → **一時停止**
84
+ 2. rule-advisor再実行
85
+ 3. 根本原因分析(なぜ?を5回繰り返して真因を特定)
86
+ 4. 対処計画提示
87
+ 5. ユーザー承認後に修正
88
+
89
+ ### エスカレーション基準(ユーザー確認必須)
90
+ - アーキテクチャ変更(新レイヤー追加、責務変更)
91
+ - 外部依存追加(npmパッケージ、外部API)
92
+ - 破壊的変更(既存APIの変更、データ構造の変更)
93
+ - 複数の実装方法があり優劣が判断できない場合
94
+
95
+ ### 集中時のルール無視防止
96
+ **測定可能な強制停止基準**:
97
+ - **連続エラー修正2回目**: 自動的にrule-advisor再実行をトリガー
98
+ - **Editツール5回使用**: 影響範囲レポートの作成を強制
99
+ - **同一ファイル3回編集**: リファクタリング検討の強制停止
100
+
101
+ ### 一時ファイル作成ルール
102
+ 作業中ファイルは`tmp/`ディレクトリ使用。完了時削除。
103
+
104
+ ### 専門エージェント積極活用(PROACTIVELY)
105
+ - **品質関連キーワード検出時**: プロジェクトタイプに応じたquality-fixer必須実行
106
+ - **Backend**: quality-fixer使用
107
+ - **Frontend**: quality-fixer-frontend使用(Lighthouse、Bundle size対応)
108
+ - 型エラー、ビルドエラー、lintエラー、フォーマット警告
109
+ - テスト失敗、品質チェック、検証タスク
110
+ - **タスク開始時**: rule-advisor必須実行
111
+ - TodoWrite作成前に実行し、結果を反映
package/LICENSE ADDED
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2025 Shinsuke Kagawa
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.