2021 | SECCON | Misc

s/<script>//gi

题目¶

Can you figure out why s/<script>//gi is insufficient for sanitizing? This can be bypassed with <scr<script>ipt>.

Remove <script> (case insensitive) from the input until the input contains no <script>.

Note that flag format is SECCON{[\x20-\x7e]+}, which means that the flag may contains < or > as the following examples.

Sample Input 1:

S3CC0N{dum<scr<script>ipt>my}

Sample Output 1:

S3CC0N{dummy}

Sample Input 2 (small.txt):

S3CC0N{dumm<scrIpT>y_flag>_<_pt>>PT><<SCr<S<<SC<SCRIpT><scRiPT>Ript>sCr<Scri<...

Sample Output 2:

S3CC0N{dummy_flag>_<_pt>>PT><sCRIp<scr<scr<scr!pt>ipt>ipt>}

解题思路¶

不断移除 <script>（大小写不敏感）标签，直到没有为止
然而 flag.txt 大小高达 67M

先用 sed 替换几下，能砍掉 20 M

for i in {1..10};do sed -i 's/<[sS][cC][rR][iI][pP][tT]>//g' flag.txt; done

然后，进入瓶颈阶段 -_- 「用啥都不会太快」（Python 和 C++ 都跑了两小时左右）

s = open('flag.bak', 'r').read()
l = len(s)

res, p = '', 0

for i in range(l):
    if s[i] == '>':
        res += s[p:i+1]
        while '<script>' == res[-8:].lower():
            res = res[:-8]
        p = i + 1

if p != l:
    res += s[p:l]        

print(res)

#include<iostream>
#include<fstream>
#include<algorithm>
using namespace std;

int main()
{
    ifstream infile;
    infile.open("flag.txt");
    string s, r = "";
    infile >> s;
    infile.close();

    int l = s.length(), p = 0;
    for (int i = 0; i < l; i++) {
        if (s[i] == '>') {
            r += s.substr(p, i - p + 1);
            int tl = r.length();
            string tmp = r.substr(tl - 8);
            transform(tmp.begin(), tmp.end(), tmp.begin(), ::tolower);
            if ("<script>" == tmp) {
                r = r.substr(0, tl - 8);
            }
            p = i + 1;
        }
    }

    if (p != l) {
        r += s.substr(p, l - p + 1);
    }

    cout << r << endl;
    return 0;
}

最终结果：SECCON{sanitizing_is_not_so_good><_escaping_is_better_iPt><SCript<ScrIpT<scRIp<scRI<Sc<scr!pt>}

最后更新: 2021年12月12日 15:57:17

Contributors: YanhuiJessica
Pageviews:

s/<script>//gi

题目¶

解题思路¶

评论