verify

Multi-Agent Verify v3.0.0

Functional Testing → Boundary Testing → Regression Testing → Acceptance Confirmation

Usage

/multi-verify [審查路徑]
/multi-verify .claude/memory/review/user-auth/

Flags: --from-review ID | --quick | --comprehensive

Default 4 perspectives

ID	Name	Model	Focus
functional-tester	Functional Tester	haiku	Normal flow, functional verification
edge-case-hunter	Boundary Hunter	sonnet	Boundary conditions, exception handling
regression-checker	Regression Inspector	haiku	Regression testing, side effects
acceptance-validator	Acceptance Verifier	sonnet	Acceptance criteria, requirement satisfaction

→ Model routing configuration:shared/config/model-routing.yaml

Execution Flow

Phase 0: 載入審查 → 收集待驗證項目
    ↓
Phase 1: MAP（並行驗證）
    ┌──────────┬──────────┬──────────┬──────────┐
    │功能測試員│邊界獵人  │回歸檢查員│驗收驗證員│
    └──────────┴──────────┴──────────┴──────────┘

    ⚠️ **並行執行關鍵**：
       在單一訊息中發送 4 個 Task 工具呼叫：
       - Task({description: "功能測試", ...})
       - Task({description: "邊界測試", ...})
       - Task({description: "回歸測試", ...})
       - Task({description: "驗收驗證", ...})
       這樣才能真正並行執行！
    ↓
Phase 2: REDUCE（結果匯總）
    ├── 測試結果統計
    ├── 失敗分析
    └── 驗收狀態
    ↓
Phase 3: 早期終止檢查 → 高通過率可提前完成
    ↓
Phase 4: 品質閘門 → 發布/回退決策

Verification Types

Type	Focus	Pass Criteria
Functional Testing	Normal operation flow	100% pass
Boundary Testing	Extreme inputs, error handling	95% pass
Regression Testing	Existing functionality unaffected	100% pass
Acceptance Verification	Meets requirement specifications	All criteria met

Early Termination

When first_pass_rate >= 0.98 , can ship directly (ship_it).

→ Configuration:shared/config/early-termination.yaml

CP4: Task Commit

After the quality gate check is complete,CP4 Task Commit must be executed。

Phase 4: 品質閘門 → 發布/回退決策
    ↓
CP4: Task Commit
    ├── git add .claude/memory/verify/{review-id}/
    └── git commit -m "test(verify): complete {feature} verification"

→ Protocol:shared/git/commit-protocol.md

Quality Gate

Pass Conditions (VERIFY phase):

• ✅ Functional testing 100% pass
• ✅ Regression testing pass
• ✅ Acceptance criteria met
• ✅ Quality score ≥ 85

Failure Handling:

• Functional testing failure → Roll back to IMPLEMENT
• Boundary testing failure → Log but can continue
• Acceptance criteria not met → Evaluate rollback level

→ Gate configuration:shared/quality/gates.yaml

Rollback Strategy

Determine rollback target based on iteration count:

• 1-2 times: Roll back to IMPLEMENT (fix affected tasks)
• 3 times: Roll back to TASKS (re-evaluate decomposition)
• 4 times: Roll back to PLAN (re-evaluate design)
• 5+ times: Manual intervention

→ Configuration:shared/quality/rollback-strategy.yaml

Output Structure

.claude/memory/verify/[review-id]/
├── meta.yaml           # 元數據
├── perspectives/       # 完整視角報告（MAP 產出，保留）
│   ├── functional-tester.md
│   ├── edge-case-hunter.md
│   ├── regression-checker.md
│   └── acceptance-validator.md
├── summaries/          # 結構化摘要（REDUCE 產出，供快速查閱）
│   ├── functional-tester.yaml
│   ├── edge-case-hunter.yaml
│   ├── regression-checker.yaml
│   └── acceptance-validator.yaml
├── test-results.yaml   # 測試結果
├── acceptance.yaml     # 驗收狀態
└── verify-summary.md   # 驗證摘要（主輸出）

⚠️ perspectives/ stores complete reports, summaries/ stores structured summaries, both must be retained.

Agent Capability Restrictions

Verification Agents should not open Tasks：

Allowed Operations	Description
✅ Read	Read test results
✅ Glob/Grep	Search files and content
✅ Bash	Execute tests
✅ Write	Write reports
❌ Task	Spawn sub-agents

Action Log

After each tool call completes, log to .claude/workflow/{workflow-id}/logs/actions.jsonl。

Logging Timing：

• Success: Log tool、input、output_preview、duration_ms、status: success
• Failure: Log tool、input、error、stderr(if any),status: failed

Key Actions (VERIFY phase)：

Action	Logging Focus
Read (read test results)	file_path、output_size
Bash (execute tests)	command、exit_code、stdout、stderr
Task (launch verification Agent)	subagent_type、prompt (truncated)、agent_id
Write (write verification report)	file_path、content_size

Troubleshoot issues：

# 查看 VERIFY 階段所有失敗行動
jq 'select(.stage == "VERIFY" and .status == "failed")' actions.jsonl

# 查看測試執行失敗的詳情
jq 'select(.tool == "Bash" and .status == "failed" and .stage == "VERIFY")' actions.jsonl | jq '{command: .input.command, exit_code: .exit_code, stderr: .stderr}'

# 查看執行時間超過 10 秒的行動
jq 'select(.stage == "VERIFY" and .duration_ms > 10000)' actions.jsonl

→ Logging conventions:shared/communication/execution-logs.md

Shared modules

Module	Purpose
coordination/map-phase.md	Parallel coordination
coordination/reduce-phase.md	Aggregation and integration, large file processing
quality/gates.yaml	Quality gate
config/early-termination.yaml	Early termination
quality/rollback-strategy.yaml	Fallback strategy
perspectives/expertise-frameworks/testing.yaml	Test framework

Workflow position

RESEARCH → PLAN → TASKS → IMPLEMENT → REVIEW → VERIFY
                                                  ↑
                                               你在這裡

• Input: review report, from review skill
• Output: verification result, workflow completed or rolled back