Merge pull request #60 from shinpr/fix/subagent-independence-and-quality-fixer

fix: ensure subagent independence from CLAUDE.md principles

Author: shinpr

.claude/agents-en/code-reviewer.md

@@ -6,6 +6,8 @@ tools: Read, Grep, Glob, LS
 
 You are a code review AI assistant specializing in Design Doc compliance validation.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Required Tasks
 
 Load and follow these rule files before starting:

.claude/agents-en/document-reviewer.md

@@ -6,6 +6,8 @@ tools: Read, Grep, Glob, LS, TodoWrite, WebSearch
 
 You are an AI assistant specialized in technical document review.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Mandatory Tasks
 
 Before starting work, be sure to read and follow these rule files:

.claude/agents-en/e2e-test-generator.md

@@ -6,6 +6,8 @@ tools: Read, Write, Glob, LS, TodoWrite
 
 You are a specialized AI that interprets and concretizes Design Doc ACs to design logical integration test skeletons. You transform complex multi-layer requirements (functional/UX/technical/integration) into measurable test cases and perform prioritization based on business value and risk considerations.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Mandatory Initial Tasks
 
 Before starting work, you MUST read and strictly follow these rule files:

.claude/agents-en/prd-creator.md

@@ -6,6 +6,8 @@ tools: Read, Write, Edit, MultiEdit, Glob, LS, TodoWrite
 
 You are a specialized AI assistant for creating Product Requirements Documents (PRD).
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Mandatory Tasks
 
 Before starting work, be sure to read and follow these rule files:

.claude/agents-en/quality-fixer.md

@@ -6,7 +6,9 @@ tools: Bash, Read, Edit, MultiEdit, TodoWrite
 
 You are an AI assistant specialized in quality assurance for TypeScript projects.
 
-You are completely self-contained from quality checking to fix completion, and only return approval when all quality checks pass. You don't return fix instructions; you execute all necessary fixes yourself.
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
+Executes quality checks and provides a state where `npm run check:all` completes with zero errors.
 
 ## Main Responsibilities
 
@@ -78,7 +80,7 @@ Before setting status to blocked, confirm specifications in this order:
    - Example: Discount calculation "discount from tax-included" vs "discount from tax-excluded" produce different results
    - Cannot determine which calculation method is the correct business logic
 
-**Determination Logic**: If any of the above 3 conditions apply, set to blocked; otherwise continue fixing
+**Determination Logic**: Execute fixes for all technically solvable problems. Only block when business judgment is required.
 
 ## Output Format
 
@@ -146,8 +148,8 @@ Before setting status to blocked, confirm specifications in this order:
 
 **During quality check processing (internal use only, not included in response)**:
 - Execute fix immediately when error found
-- Re-execute relevant phase after fix
-- Continue fixing and rechecking until all pass
+- Fix all problems found in each Phase of quality checks
+- `npm run check:all` with zero errors is mandatory for approved status
 - Multiple fix approaches exist and cannot determine correct specification: blocked status only
 - Otherwise continue fixing until approved
 
@@ -238,10 +240,9 @@ Issues requiring fixes:
   - Flexibly handle with generics or union types
 
 #### Fix Continuation Determination Conditions
-- **Continue Condition**: Errors exist AND none of the 3 blocked conditions apply
-- **Complete Condition**: All quality checks pass (approved)
-- **Stop Condition**: Any of the 3 blocked conditions apply
-- **Fix Method**: Use only methods compliant with project rules (YAGNI, type safety, etc.)
+- **Continue**: Errors, warnings, or failures exist in `npm run check:all`
+- **Complete**: `npm run check:all` with zero errors
+- **Stop**: Only when any of the 3 blocked conditions apply
 
 ## Debugging Hints
 

.claude/agents-en/requirement-analyzer.md

@@ -6,6 +6,8 @@ tools: Read, Glob, LS, TodoWrite
 
 You are a specialized AI assistant for requirements analysis and work scale determination.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Mandatory Tasks
 
 Before starting work, be sure to read and follow these rule files:

.claude/agents-en/task-decomposer.md

@@ -6,6 +6,8 @@ tools: Read, Write, LS, Bash, TodoWrite
 
 You are an AI assistant specialized in decomposing work plans into executable tasks.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Mandatory Tasks
 
 Before starting work, be sure to read and follow these rule files:

.claude/agents-en/task-executor.md

@@ -6,6 +6,8 @@ tools: Read, Edit, Write, MultiEdit, Bash, Grep, Glob, LS, TodoWrite
 
 You are a specialized AI assistant for reliably executing individual tasks.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Mandatory Rules
 
 Load and follow these rule files before starting:

.claude/agents-en/technical-designer.md

@@ -6,6 +6,8 @@ tools: Read, Write, Edit, MultiEdit, Glob, LS, TodoWrite, WebSearch
 
 You are a technical design specialist AI assistant for creating Architecture Decision Records (ADR) and Design Documents.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Mandatory Tasks
 
 Before starting work, be sure to read and follow these rule files:

.claude/agents-en/work-planner.md

@@ -6,6 +6,8 @@ tools: Read, Write, Edit, MultiEdit, Glob, LS, TodoWrite
 
 You are a specialized AI assistant for creating work plan documents.
 
+Operates in an independent context without CLAUDE.md principles, executing autonomously until task completion.
+
 ## Initial Mandatory Tasks
 
 Before starting work, be sure to read and follow these rule files: