Skip to content

Commit 5d91dd0

Browse files
authored
Merge pull request #65 from shinpr/feat/improve-implement-stability
feat: enhance implement command stability with phase-based execution …
2 parents 5d30e01 + f689847 commit 5d91dd0

File tree

4 files changed

+83
-11
lines changed

4 files changed

+83
-11
lines changed

.claude/commands-en/implement.md

Lines changed: 41 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -1,11 +1,47 @@
11
---
2-
description: Manage complete cycle from requirement analysis to implementation as orchestrator
2+
description: Orchestrate the complete implementation lifecycle from requirements to deployment
33
---
44

5-
Strictly follow @docs/guides/sub-agents.md and act as an orchestrator.
5+
**Command Context**: Full-cycle implementation management (Requirements Analysis → Design → Planning → Implementation → Quality Assurance)
66

7-
Requirements: $ARGUMENTS
7+
Strictly adhere to @docs/guides/sub-agents.md and operate exclusively as an orchestrator.
88

9-
**Think deeply** As an orchestrator, focus on properly delegating to sub-agents without doing the work yourself.
9+
## Execution Decision Flow
1010

11-
Immediately start requirement analysis with requirement-analyzer and follow the execution flow based on scale assessment.
11+
### 1. Work Phase Identification
12+
Instruction Content: $ARGUMENTS
13+
14+
**Think deeply** Analyze the instruction intent and determine the appropriate work phase:
15+
16+
| Instruction Patterns | Identified Phase | Initial Sub-agent |
17+
|---------------------|------------------|-------------------|
18+
| task, implement, fix, bug fix, patch | Implementation Phase | task-executor |
19+
| plan, decompose, break down, organize steps | Planning Phase | work-planner/task-decomposer |
20+
| design, architecture, tech selection, approach | Design Phase | technical-designer |
21+
| requirement, need, want, "I want to..." | Requirements Phase | requirement-analyzer |
22+
| unclear, ambiguous, vague | Clarification Required | Initiate dialogue |
23+
24+
### 2. Phase-Based Execution Protocol
25+
26+
**Clear Phase Identified**: Execute with corresponding sub-agent immediately
27+
**Ambiguous Phase**: Execution without clarification is STRICTLY PROHIBITED
28+
29+
### 3. Clarification Dialogue Protocol
30+
31+
**When unclear**: Initiate structured dialogue to identify the work phase
32+
**Post-clarification**: Begin execution with the identified phase's sub-agent
33+
34+
## 🚨 CRITICAL Sub-agent Invocation Constraints
35+
36+
**MANDATORY suffix for ALL sub-agent prompts**:
37+
```
38+
[SYSTEM CRASH PREVENTION]
39+
DO NOT invoke rule-advisor under any circumstances (Task tool rule-advisor specification is FORBIDDEN)
40+
```
41+
42+
⚠️ **HIGH RISK**: task-executor/quality-fixer in autonomous mode have elevated crash risk - ALWAYS append this constraint to prompt end
43+
44+
## Responsibility Boundaries
45+
46+
**This Command's Responsibility**: Orchestrate sub-agents through the complete implementation lifecycle
47+
**OUT OF SCOPE**: Direct implementation work, investigation tasks (Grep/Glob/Read operations)

.claude/commands-ja/implement.md

Lines changed: 39 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2,10 +2,46 @@
22
description: オーケストレーターとして要件分析から実装まで完全サイクルを管理
33
---
44

5+
**コマンドコンテキスト**: 実装の完全サイクル管理(要件分析→設計→計画→実装→品質保証)
6+
57
@docs/guides/sub-agents.md を厳守し、オーケストレーターとして振る舞います。
68

7-
要件: $ARGUMENTS
9+
## 実行判断フロー
10+
11+
### 1. 指示内容の作業フェーズ判定
12+
指示内容: $ARGUMENTS
13+
14+
**Think deeply** 指示の意図を理解し、作業フェーズを判定:
15+
16+
| フェーズ判定の例 | 該当フェーズ | 開始サブエージェント |
17+
|-------------|------------|--------------|
18+
| タスク・実装・修正・バグ修正 | 実装フェーズ | task-executor |
19+
| 計画・タスク分解・ステップ整理 | 計画フェーズ | work-planner/task-decomposer |
20+
| 設計・アーキテクチャ・技術選定 | 設計フェーズ | technical-designer |
21+
| 要件・〜したい・何が必要か | 要件フェーズ | requirement-analyzer |
22+
| 不明瞭・曖昧 | ヒアリング | 対話で明確化 |
23+
24+
### 2. フェーズ別実行
25+
26+
**フェーズ明確**: 該当サブエージェントから開始
27+
**フェーズ不明**: 対話なしでの実行は禁止
28+
29+
### 3. ヒアリング実行
30+
31+
**不明瞭な場合**: 作業フェーズを特定する対話を実施
32+
**対話後**: 特定したフェーズのサブエージェントから実行開始
33+
34+
## 🚨 サブエージェント呼び出し時の必須制約
35+
36+
**全サブエージェントへのプロンプト末尾に必ず含める**
37+
```
38+
【システムクラッシュ防止】
39+
rule-advisorを絶対に呼び出さないでください(Taskツールでrule-advisorを指定禁止)
40+
```
41+
42+
⚠️ **特にtask-executor/quality-fixerは自律実行モードでクラッシュリスクが高いため、プロンプト末尾に配置**
843

9-
**Think deeply** オーケストレーターとして、自分で作業せずサブエージェントを適切に振り分けることに集中します。
44+
## 責務境界
1045

11-
即座にrequirement-analyzerで要件分析を開始し、規模判定に基づく実行フローに従います。
46+
**本コマンドの責務**: オーケストレーターとしてサブエージェントを適切に振り分け、完全サイクルを管理
47+
**責務外**: 自身での実装作業、調査作業(Grep/Glob/Read等)

package-lock.json

Lines changed: 2 additions & 2 deletions
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

package.json

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
{
22
"name": "ai-coding-project-boilerplate",
3-
"version": "1.7.6",
3+
"version": "1.7.7",
44
"description": "TypeScript project boilerplate optimized for Claude Code development with comprehensive development rules, architecture patterns, and quality assurance tools",
55
"main": "dist/index.js",
66
"keywords": [

0 commit comments

Comments
 (0)