Update README table of contents for ch03 rewrite (all three languages)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: fqzz2000

en/README.md

@@ -73,13 +73,10 @@ The book unfolds along a productivity ladder. Chapter 1 analyzes the structural
   * [Conveying Intent Through Structure: Layers and Dimensions](chapters/02b-structured-intent.md)
   * [Iterating Toward an Executable Spec](chapters/02c-iterative-spec.md)
   * [In Practice: AILock-Step Feature Workflow](chapters/02d-case-study.md)
-* [The Backbone of Reliable Delivery: Automated Quality Control](chapters/03-verification.md)
-  * [From Wish List to Executable Constraints: Test First](chapters/03a-test-first.md)
-  * [AI Writes Code and Tests: The Collusion Problem](chapters/03b-collusion.md)
-  * [What Kind of Tests Are Hardest to Game: The Trophy Testing Model](chapters/03c-trophy-testing.md)
-  * [Not Check After Done, but Check While Doing: Continuous Feedback During Execution](chapters/03d-continuous-feedback.md)
-  * [Output Too Fast to Review: Using Agents to Verify Agents](chapters/03e-adversarial-verification.md)
-  * [So What Do I Do: From Reviewing Code to Designing Verification Systems](chapters/03f-new-role.md)
+* [Verification: Ensuring Code Stays Faithful to the Spec](chapters/03-verification.md)
+  * [Test Infrastructure First: Turning Specs into Executable Constraints](chapters/03a-test-first.md)
+  * [Code Review: Catching Intent Drift That Tests Miss](chapters/03b-code-review.md)
+  * [In Practice: AILock-Step's Verification Pipeline](chapters/03c-practice.md)
 
 ### Part II: Scaling Agent Development (10→100x)
 

ja/README.md

@@ -73,13 +73,10 @@ Vibe Codingは出発点だ。フィーリングでプロンプトを書き、AI
   * [構造で意図を伝える：階層化と次元](chapters/02b-structured-intent.md)
   * [実行可能なSpecをイテレーションで作る](chapters/02c-iterative-spec.md)
   * [実践：AILock-Step Feature Workflow](chapters/02d-case-study.md)
-* [信頼性あるデリバリーのバックボーン: 自動品質管理](chapters/03-verification.md)
-  * [ウィッシュリストから実行可能な制約へ: Test First](chapters/03a-test-first.md)
-  * [AIがコードもテストも書く: 共謀問題](chapters/03b-collusion.md)
-  * [どのテストが最もごまかしにくいか: Trophy Testingモデル](chapters/03c-trophy-testing.md)
-  * [終わってから確認ではなく、やりながら確認: 実行中の継続的フィードバック](chapters/03d-continuous-feedback.md)
-  * [出力が速すぎてレビューが追いつかない: AgentでAgentを検証する](chapters/03e-adversarial-verification.md)
-  * [では何をすべきか: コードレビューから検証システム設計へ](chapters/03f-new-role.md)
+* [検証：コードが仕様に忠実であることを確保する](chapters/03-verification.md)
+  * [テスト基盤の前倒し：Specを実行可能な制約に変える](chapters/03a-test-first.md)
+  * [Code Review：テストでは捕捉できない意図ドリフトを補う](chapters/03b-code-review.md)
+  * [実践：AILock-Stepの検証パイプライン](chapters/03c-practice.md)
 
 ### Part II: Agent開発のスケーリング（10→100倍）
 

zh/README.md

@@ -73,13 +73,10 @@ Vibe Coding 是起点：凭感觉写 prompt，让 AI 生成代码，能跑就行
   * [用结构传达意图：分层与维度](chapters/02b-structured-intent.md)
   * [迭代出一份可执行的 Spec](chapters/02c-iterative-spec.md)
   * [实践：AILock-Step Feature Workflow](chapters/02d-case-study.md)
-* [可靠交付的保障：自动化质量控制](chapters/03-verification.md)
-  * [从愿望清单到可执行约束：测试先行](chapters/03a-test-first.md)
-  * [AI 写代码又写测试：共谋问题](chapters/03b-collusion.md)
-  * [什么样的测试最难被作弊：Trophy 测试模型](chapters/03c-trophy-testing.md)
-  * [不是写完再查，是边写边查：执行过程中的持续反馈](chapters/03d-continuous-feedback.md)
-  * [产出速度太快验不过来：用 Agent 验 Agent](chapters/03e-adversarial-verification.md)
-  * [那我干什么：从审查代码到设计验证体系](chapters/03f-new-role.md)
+* [验证：确保代码忠实于规约](chapters/03-verification.md)
+  * [测试基建前置：把 Spec 变成可执行约束](chapters/03a-test-first.md)
+  * [Code Review：补位测试覆盖不到的意图漂移](chapters/03b-code-review.md)
+  * [实践：AILock-Step 的验证链路](chapters/03c-practice.md)
 
 ### 卷二：规模化 Agent 开发 (10→100x)