fix: pipeline LLM 全失败时明确报错而非静默输出空 IR - Closes #15

- step1: 所有 LLM 调用返回空 function_units 时抛出 RuntimeError
- step1: main() 在 _quick_validate 未通过时 sys.exit(1)
- step2: function_units 为空时提前报错终止
- step3: fragments 为空时提前报错终止
- test: test_step1 捕获 SystemExit, test_step2_5/step3 空数据改为 skip

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

skills/ir_generation_skill/step3_merge_and_audit.py

@@ -987,10 +987,17 @@ def main():
     semantic_index = load_semantic_index()
     path_enum = load_path_enumeration()
 
+    total_fragments = len(fragments)
+    if total_fragments == 0 and not autocomplete_fragments:
+        print("错误: 无 IR 片段可合并 (fragments 和 autocomplete_fragments 均为空)。")
+        print("  请检查 step2_ir_extraction 是否正确运行。")
+        print("  可能原因: step1 未生成 function_units，或 step2 提取失败。")
+        sys.exit(1)
+
     feature_name = semantic_index.get("feature_name", "行车娱乐限制")
     feature_id = "DRL-001"
     print(f"  功能: {feature_name} ({feature_id})")
-    print(f"  主片段: {len(fragments)}")
+    print(f"  主片段: {total_fragments}")
     if autocomplete_fragments:
         print(f"  自动补全片段: {len(autocomplete_fragments)}")