docs: 更新 test-cycle-execute 和 test-fix-gen 文档，增强命令范围和责任边界的描述，明确测试失败处理流程

.claude/commands/workflow/test-cycle-execute.md

@@ -10,6 +10,12 @@ allowed-tools: SlashCommand(*), TodoWrite(*), Read(*), Bash(*), Task(*)
 ## Overview
 Orchestrates dynamic test-fix workflow execution through iterative cycles of testing, analysis, and fixing. **Unlike standard execute, this command dynamically generates intermediate tasks** during execution based on test results and CLI analysis, enabling adaptive problem-solving.
 
+**⚠️ CRITICAL - Orchestrator Boundary**:
+- This command is the **ONLY place** where test failures are handled
+- All CLI analysis (Gemini/Qwen), fix task generation (IMPL-fix-N.json), and iteration management happen HERE
+- Agents (@test-fix-agent) only execute single tasks and return results
+- **Do NOT handle test failures in main workflow or other commands** - always delegate to this orchestrator
+
 **Resume Mode**: When called with `--resume-session` flag, skips discovery and continues from interruption point.
 
 ## Core Philosophy
@@ -53,7 +59,7 @@ Orchestrates dynamic test-fix workflow execution through iterative cycles of tes
 
 ## Responsibility Matrix
 
-**Clear division of labor between orchestrator and agents:**
+**⚠️ CRITICAL - Clear division of labor between orchestrator and agents:**
 
 | Responsibility | test-cycle-execute (Orchestrator) | @test-fix-agent (Executor) |
 |----------------|----------------------------|---------------------------|
@@ -62,12 +68,14 @@ Orchestrates dynamic test-fix workflow execution through iterative cycles of tes
 | Generate IMPL-fix-N.json | ✅ Creates task files | ❌ Not involved |
 | Run tests | ❌ Delegates to agent | ✅ Executes test command |
 | Apply fixes | ❌ Delegates to agent | ✅ Modifies code |
-| Detect test failures | ✅ Analyzes agent output | ✅ Reports results |
+| Detect test failures | ✅ Analyzes results and decides next action | ✅ Executes tests and reports outcomes |
 | Add tasks to queue | ✅ Manages queue | ❌ Not involved |
-| Update iteration state | ✅ Maintains state files | ✅ Updates task status |
+| Update iteration state | ✅ Maintains overall iteration state | ✅ Updates individual task status only |
 
 **Key Principle**: Orchestrator manages the "what" and "when"; agents execute the "how".
 
+**⚠️ ENFORCEMENT**: If test failures occur outside this orchestrator, do NOT handle them inline - always call `/workflow:test-cycle-execute` instead.
+
 ## Execution Lifecycle
 
 ### Phase 1: Discovery & Initialization
@@ -217,13 +225,15 @@ Iteration N (managed by test-cycle-execute orchestrator):
 **Orchestrator executes CLI analysis between agent tasks:**
 
 #### When Test Failures Occur
-1. **[Orchestrator]** Detects failures from agent output
+1. **[Orchestrator]** Detects failures from agent test execution output
 2. **[Orchestrator]** Collects failure context from `.process/test-results.json` and logs
-3. **[Orchestrator]** Runs Gemini/Qwen CLI with failure context
+3. **[Orchestrator]** Executes Gemini/Qwen CLI tool with failure context
-4. **[CLI Tool]** Analyzes failures and generates fix strategy
+4. **[Orchestrator]** Interprets CLI tool output to extract fix strategy
 5. **[Orchestrator]** Saves analysis to `.process/iteration-N-analysis.md`
 6. **[Orchestrator]** Generates `IMPL-fix-N.json` with strategy content (not just path)
 
+**Note**: The orchestrator executes CLI analysis tools and processes their output. CLI tools provide analysis, orchestrator manages the workflow.
+
 #### CLI Analysis Command (executed by orchestrator)
 ```bash
 cd {project_root} && gemini -p "
@@ -516,15 +526,16 @@ Task(subagent_type="{meta.agent}",
      ### For test-fix (IMPL-002):
      - Run test suite: {test_command}
      - Collect results to .process/test-results.json
-     - If failures: Save context, return to orchestrator
+     - Report results to orchestrator (do NOT analyze failures)
+     - Orchestrator will handle failure detection and iteration decisions
      - If success: Mark complete
 
      ### For test-fix-iteration (IMPL-fix-N):
      - Load fix strategy from context.fix_strategy (CONTENT, not path)
      - Apply surgical fixes to identified files
-     - Run tests to verify
+     - Return results to orchestrator
-     - If still failures: Save context with new failure data
+     - Do NOT run tests independently - orchestrator manages all test execution
-     - Update iteration state
+     - Do NOT handle failures - orchestrator analyzes and decides next iteration
 
      ## STEP 4: Implementation Context (From JSON)
      **Requirements**: {context.requirements}

.claude/commands/workflow/test-fix-gen.md

@@ -13,7 +13,11 @@ allowed-tools: SlashCommand(*), TodoWrite(*), Read(*), Bash(*)
 
 This command creates an independent test-fix workflow session for existing code. It orchestrates a 5-phase process to analyze implementation, generate test requirements, and create executable test generation and fix tasks.
 
-**⚠️ Command Scope**: Prepares test workflow artifacts only. Task execution requires separate commands (`/workflow:test-cycle-execute` or `/workflow:execute`).
+**⚠️ CRITICAL - Command Scope**:
+- **This command ONLY generates task JSON files** (IMPL-001.json, IMPL-002.json)
+- **Does NOT execute tests or apply fixes** - all execution happens in separate orchestrator
+- **Must call `/workflow:test-cycle-execute`** after this command to actually run tests and fixes
+- **Test failure handling happens in test-cycle-execute**, not here
 
 ### Dual-Mode Support
 
@@ -44,12 +48,15 @@ fi
 
 ### Coordinator Role
 
-This command is a **pure orchestrator**:
+This command is a **pure planning coordinator**:
 - Does NOT analyze code directly
 - Does NOT generate tests or documentation
-- ONLY coordinates slash commands in sequence
+- Does NOT execute tests or apply fixes
+- Does NOT handle test failures or iterations
+- ONLY coordinates slash commands to generate task JSON files
 - Parses outputs to pass data between phases
 - Creates independent test workflow session
+- **All execution delegated to `/workflow:test-cycle-execute`**
 
 ---
 
@@ -267,14 +274,20 @@ Review artifacts:
 - Test plan: .workflow/[testSessionId]/IMPL_PLAN.md
 - Task list: .workflow/[testSessionId]/TODO_LIST.md
 
-Next Steps:
+⚠️ CRITICAL - Next Steps:
-- Review IMPL_PLAN.md
+1. Review IMPL_PLAN.md
-- Execute: /workflow:test-cycle-execute [testSessionId]
+2. **MUST execute: /workflow:test-cycle-execute**
+   - This command only generated task JSON files
+   - Test execution and fix iterations happen in test-cycle-execute
+   - Do NOT attempt to run tests or fixes in main workflow
 ```
 
 **TodoWrite**: Mark phase 5 completed
 
-**Note**: Command completes here. Task execution requires separate workflow commands.
+**⚠️ BOUNDARY NOTE**:
+- Command completes here - only task JSON files generated
+- All test execution, failure detection, CLI analysis, fix generation happens in `/workflow:test-cycle-execute`
+- This command does NOT handle test failures or apply fixes
 
 ---
 
@@ -329,7 +342,9 @@ Generates minimum 2 tasks (expandable for complex projects):
 
 **Agent**: `@test-fix-agent`
 
-**Purpose**: Execute tests and apply iterative fixes (max 5 iterations)
+**Purpose**: Execute initial tests and trigger orchestrator-managed fix cycles
+
+**Note**: This task executes tests and reports results. The test-cycle-execute orchestrator manages all fix iterations, CLI analysis, and fix task generation.
 
 **Task Configuration**:
 - Task ID: `IMPL-002`
@@ -340,11 +355,12 @@ Generates minimum 2 tasks (expandable for complex projects):
 - `context.requirements`: Execute and fix tests
 
 **Test-Fix Cycle Specification**:
-- **Cycle Pattern**: test → gemini_diagnose → manual_fix (or codex) → retest
+**Note**: This specification describes what test-cycle-execute orchestrator will do. The agent only executes single tasks.
-- **Tools Configuration**:
+- **Cycle Pattern** (orchestrator-managed): test → gemini_diagnose → manual_fix (or codex) → retest
+- **Tools Configuration** (orchestrator-controlled):
   - Gemini for analysis with bug-fix template → surgical fix suggestions
   - Manual fix application (default) OR Codex if `--use-codex` flag (resume mechanism)
-- **Exit Conditions**:
+- **Exit Conditions** (orchestrator-enforced):
   - Success: All tests pass
   - Failure: Max iterations reached (5)