fix: read_file支持解析docx(zipfile提取文本),避免agent读docx报错绕去run_command探测

pipeline_service/agent_loop_v2.py

@@ -775,8 +775,21 @@ class AgentExecutor:
         try:
             if not os.path.isfile(full):
                 return f"FAIL: 文件不存在 {path}"
-            with open(full, encoding="utf-8") as f:
+            ext = os.path.splitext(full)[1].lower()
-                return f.read()[:30000]
+            # docx：提取 word/document.xml 文本
+            if ext == '.docx':
+                import zipfile
+                import re as _re
+                with zipfile.ZipFile(full) as z:
+                    xml = z.read('word/document.xml').decode('utf-8', errors='ignore')
+                    texts = _re.findall(r'<w:t[^>]*>(.*?)</w:t>', xml)
+                    return ('\n'.join(texts)[:30000]) or '(docx 无文本内容)'
+            # 纯文本类：直接读
+            if ext in ('.txt', '.md', '.json', '.csv', '.py', '.log', '.yaml', '.yml', '.xml', '.html', '.ini', ''):
+                with open(full, encoding="utf-8", errors="ignore") as f:
+                    return f.read()[:30000]
+            # 其他二进制
+            return f"该文件是二进制格式({ext or '无扩展名'})，无法直接读取文本。可改用 run_command 处理，或让用户上传文本版本。"
         except Exception as e:
             return f"ERROR: {str(e)[:300]}"