docs: Phase 109 DONE

- Create phases/phase-109/README.md
- Update 10-Now.md with completion report

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

docs/development/current/main/10-Now.md

@@ -1,5 +1,13 @@
 # Self Current Task — Now (main)
 
+## 2025-12-17：Phase 109 完了 ✅
+
+**Phase 109: error_tags hints SSOT**
+- policy/validator エラーを "tag + message + hint" 形式に統一
+- freeze_with_hint() API 追加（hint 必須、空なら panic）
+- Phase 107/104/100 の代表3箱を hint 対応に移行
+- 入口: `docs/development/current/main/phases/phase-109/README.md`
+
 ## 2025-12-17：Phase 104 完了 ✅
 
 **Phase 104: loop(true) + break-only digits（read_digits 系）**
@@ -20,6 +28,13 @@
   - LLVM EXE: `tools/smokes/v2/profiles/integration/apps/phase107_find_balanced_array_end_llvm_exe.sh` / `tools/smokes/v2/profiles/integration/apps/phase107_find_balanced_object_end_llvm_exe.sh`
 - 入口: `docs/development/current/main/phases/phase-107/README.md`
 
+## 2025-12-17：Phase 108 完了 ✅
+
+**Phase 108: Pattern2 policy router SSOT（薄い入口の固定）**
+- post-loop early return を一般 plan（`cond`/`ret_expr`）として独立し、Pattern2Inputs の依存を解消
+- ApplyPolicyStepBox から “直叩き” を撤去し、policy router 1本に集約（入口SSOT）
+- 入口: `docs/development/current/main/phases/phase-108/README.md`
+
 ## 2025-12-17：Phase 103 P0 完了 ✅
 
 **Phase 103: if-only regression baseline**

docs/development/current/main/phases/phase-109/README.md

@@ -0,0 +1,5 @@
+# Phase 109: error_tags hints SSOT
+
+**Purpose**: Unify policy/validator errors with "tag + message + hint" format.
+**Changes**: Added `freeze_with_hint()` API, migrated Phase 107/104/100 representative policies.
+**Acceptance**: cargo test --lib + smoke regression PASS.