首页
社区
课程
招聘
[原创]CTF之自动化VM分析
发表于: 2021-9-29 21:41 12727

[原创]CTF之自动化VM分析

2021-9-29 21:41
12727

CTF之自动化VM

案例:2021长安杯-SimpleVM
一道当时比赛零解的题
图片描述
图片描述
题目文件以及脚本文件见附件

一:程序分析

拖入ida知道输入长度为8

 

输入flag{11223344}进行调试
发现将flag的内容11223344分开存放到两个四字节的变量中

 

然后将其移动到了内存中的某个位置(vm结构体中的寄存器)

 

下图是加载vm的bytecode

 

图片描述

 

将flag的内容移动到vm结构体中
图片描述

 

然后下面一个函数就是dispatcher,典型的用while循环读取的字节码进行分发执行不同的虚拟指令

 

再下方就是检测两个寄存器是否是0,为0才right
图片描述

二:VM分析

分析VM,先找到它的bytecode,我们可以在内存中找到

 

图片描述

 

然后找dispatcher

 

图片描述

 

然后就是分析如何模拟的指令,以及根据这些模拟的指令来分析vm结构体

 

我恢复出来的vm结构体:

 

根据内存进行分析得到的

1
2
3
4
5
6
7
8
9
10
11
struct vm
{
    qword rip;
    qword* opcode;
    qword opcode_length;
    dword regs0;
    dword regs1;
        dword regs2;
        dword regs3;
        .....
}

然后就是分析虚拟机的模拟指令

1.每次循环获取opcode

get_rip_value_ripad1函数:获取rip地址对应的字节码,然后将rip+1

 

图片描述

2. mov regs[num1], regs[num2]

图片描述

 

space首地址偏移24字节开始是存放的寄存器

 

所以这条指令解析为:

 

mov regs[操作数1], regs[操作数2]

3. add regs[num1], regs[num2]

图片描述
同上,不过这里是+=

 

所以这条指令解析为:

 

add regs[操作数1],regs[操作数2]

4. shl regs[num1], regs[num2]

图片描述

 

同上,解析为:

 

shl regs[操作数1], regs[操作数2]

5. shr regs[num1], regs[num2]

图片描述

 

同上,解析为:

 

shr regs[num1], regs[num2]

6. regs[num1] %= regs[num2]

图片描述

 

同上,解析为:

 

regs[num1] %= regs[num2]

7. xor regs[num1], regs[num2]

图片描述

8. mov regs[num1], [regs[num2]+0x86]

将内存中的值赋值给寄存器

 

图片描述

9. mov [regs[num1]+0x86], regs[num2]

图片描述

10. cmp regs[num1], regs[num2]

比较两个寄存器并设置标志位

 

图片描述

11. nop

图片描述
图片描述

 

其实这里是设计的一个简单的花指令,获取操作数之后

 

rip += 操作数

 

但是后来发现,基本上都是+1,显然是插了一个字节

 

直接nop

12. 跳转

图片描述

三:编写反汇编器

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
class Decompile:
    def __init__(self, file_name, opcode_keys_, code):
        self.filename = file_name   # 要逆向的文件路径
        self.asm_code = []          # 反汇编的指令列表
        self.pc = 0                 # 虚拟机的rip
        self.code = code            # 存放文件的虚拟code
        self.opcode_keys = opcode_keys_  # 存放文件的指令集代表的opcode
        self.recompile_to_c = ""    # 二维数组来存放每个文件反编译成c的代码
        self.key_table = []         # 二维数组来存放文件的加密用的key_table
        self.enc_number = 0         # 根据key来判断加密的次数
        self.a0 = 0                 # 加密之后的数据(后面得到key_table之后会设置)
        self.a1 = 0
        self.i = 0                 # 解密的递归循环变量,用来判断是否退出递归
        self.string = ''           # 解密之后的字符串
    def disamone(self):          # 反编译为伪汇编指令
        regs = ['eax', 'ebx', 'ecx', 'edx', 'esi', 'edi', 'reg6', 'reg7', 'reg8', 'reg9', 'reg10', 'reg11', 'reg12', 'reg12', 'reg14', 'reg15', 'reg16', 'reg17', 'reg18', 'reg19', 'eax', 'ebx', 'reg22', 'reg23', 'reg24', 'reg25', 'reg26', 'reg27', 'reg28', 'reg29', 'ecx', 'edx', 'reg32']
        opcode = self.get_opcode()
        offset = len(self.code)
        if opcode == self.opcode_keys[11]:
            pc = self.pc
            num1,num2 = self.rip_ad_six()
            self.asm_code.append("_%s: mov %s, %s" % (hex(offset+pc), regs[num1], num2))
        elif opcode == self.opcode_keys[0]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov %s, %s" % (hex(offset+pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[1]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: add %s, %s" % (hex(offset+pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[2]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: shl ecx, al" % (hex(offset + pc)))        # regs[num2] % (hex(offset + pc), regs[num1])
        elif opcode == self.opcode_keys[3]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: shr edx, bl" % (hex(offset + pc)))       # , regs[num2] % (hex(offset + pc), regs[num1])
        elif opcode == self.opcode_keys[4]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov eax, %s\ncdq\nidiv %s\nmov %s, edx" % (hex(offset + pc), regs[num1], regs[num2], regs[num1]))
        elif opcode == self.opcode_keys[7]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: xor %s, %s" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[5]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov %s, [%s+0x86]" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[6]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov [%s+0x86], %s" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[8]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: cmp %s, %s" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[9]:
            pc = self.pc
            num1 = self.rip_ad_five()
            self.asm_code.append("_%s: nop" % (hex(offset + pc)))
        elif opcode == self.opcode_keys[10]:
            pc = self.pc
            num1, val = self.used_by_jcc()
            jcc = ['jmp', 'jl', 'jg']
            temp = ''
            if val == 0:
                temp = jcc[0]
            elif val == 1:
                temp = jcc[1]
            elif val == 2:
                temp = jcc[2]
            self.asm_code.append("_%s: %s %s" % (hex(offset + pc), temp, num1))
        else:
            print("undefined instruction")
        def get_asm_code(self):         # 获取伪汇编代码
        while self.pc < len(self.code):
            self.disamone()
        print("\n".join(self.asm_code))

得到伪汇编指令为:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
_0x37f8: mov eax, 4
_0x37fe: mov ebx, 5
_0x3804: mov ecx, ebx
_0x3807: mov edx, ebx
_0x380a: shl ecx, al
_0x380d: nop
_0x3813: shr edx, bl
_0x3816: mov esi, ecx
_0x3819: nop
_0x381f: xor esi, edx
_0x3822: add esi, ebx
_0x3825: mov edi, 64
_0x382b: xor esi, edi
_0x382e: add eax, esi
_0x3831: mov ecx, eax
_0x3834: nop
_0x383a: mov edx, eax
_0x383d: nop
_0x3843: shl ecx, al
_0x3846: shr edx, bl
_0x3849: nop
_0x384f: mov esi, ecx
_0x3852: xor esi, edx
.........

asm.txt

四:优化汇编利用pwntools的asm直接将指令转换成机器码

缺点:对汇编的格式要求很高,必须要严格满足对应架构

 

有些指令在写反汇编器的时候很难实现,比如乘,除等

 

对于这道题,想了一下,并没有涉及到操作栈的指令,几乎全部是用寄存器来操作的

 

所以,直接重编译成C

五:重编译成C直接导入编译器编译

寄存器直接声明成变量即可

 

初始化工作:

 

用变量代替寄存器,用数组代替虚拟机内存,声明变量看作eflag

 

图片描述

 

然后编写重编译器

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
def disamtoc(self):                # 根据分析反编为C代码
        regs = ['a0', 'a1', 'a2', 'a3', 'a4', 'a5', 'a6', 'a7', 'a8', 'a9', 'a10', 'a11', 'a12', 'a13', 'a14', 'a15', 'a16', 'a17', 'a18', 'a19', 'a20', 'a21', 'a22', 'a23', 'a24', 'a25', 'a26', 'a27', 'a28', 'a29', 'a30', 'a31', 'a32']
        opcode = self.get_opcode()
        if opcode == self.opcode_keys[11]:
            pc = self.pc
            num1, num2 = self.rip_ad_six()
            # self.recompile_to_c += "_{}: {}={};\n".format(hex(pc), regs[num1], num2)
            self.recompile_to_c += "{}={};\n".format(regs[num1], num2)
        elif opcode == self.opcode_keys[0]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}={};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[1]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}+{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}+={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[2]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}<<{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}<<={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[3]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}>>{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}>>={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[4]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {} % {};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{} %= {};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[7]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}^{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{} ^= {};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[5]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}=arr[{}+0x86];\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}=arr[{}+0x86];\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[6]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: arr[{}+0x86]={};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "arr[{}+0x86]={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[8]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.recompile_to_c += "eflag={}-{};\n".format(regs[num1], regs[num2])
            # self.recompile_to_c += "_{}: eflag={}-{};\n".format(hex(pc), regs[num1], regs[num2])
        elif opcode == self.opcode_keys[9]:
            pc = self.pc
            num1 = self.rip_ad_five()
            # self.recompile_to_c += "_{}: goto _{}".format(hex(pc), hex(pc+1))
            self.recompile_to_c += "\n"
        elif opcode == self.opcode_keys[10]:
            pc = self.pc
            num1, val = self.used_by_jcc()
            jcc = ['jmp', 'jl', 'jg']
            temp = ''
            if val == 0:
                temp = jcc[0]
            elif val == 1:
                temp = jcc[1]
            elif val == 2:
                temp = jcc[2]
            self.recompile_to_c += "\n"
            # self.recompile_to_c += "_{}: val={};\nif(val == 0):goto _{};\nif(val==1 && eflag<0):goto _{}\nif(val==2 && eflag>0):goto _{}\n".format(hex(pc), val, hex(pc+num1), hex(pc+num1), hex(pc+num1))
        else:
            print("undefined instruction")
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
a20=0x00000004;
a21=0x00000005;
a2=a1;
a3=a1;
a2<<=a20;
 
a3>>=a21;
a4=a2;
 
a4^= a3;
a4+=a1;
a5=0x7b7cc140;
a4^= a5;
a0+=a4;
a2=a0;
 
a3=a0;
 
a2<<=a20;
a3>>=a21;
 
a4=a2;
a4^= a3;
a4+=a0;
a5=0x2105b8c9;
a4^= a5;
a1+=a4;
a20=0x00000004;
a21=0x00000005;
 
a2=a1;
a3=a1;
a2<<=a20;
a3>>=a21;
a4=a2;
 
a4^= a3;
 
a4+=a1;
 
a5=0xfea4fb66;
a4^= a5;
 
a0+=a4;
a2=a0;
a3=a0;
a2<<=a20;
 
a3>>=a21;
a4=a2;
a4^= a3;
 
a4+=a0;
a5=0x7d1ae6c0;
a4^= a5;
a1+=a4;
a20=0x00000004;
 
a21=0x00000005;
a2=a1;
a3=a1;
a2<<=a20;
a3>>=a21;
 
a4=a2;
a4^= a3;
 
a4+=a1;
a5=0x2775177e;
a4^= a5;
a0+=a4;
 
......

然后复制到编译器中直接进行编译

 

得到程序的逻辑

 

编译后将exe拖入ida

 

图片描述

 

可以得到程序的逻辑

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
int __cdecl main(int argc, const char **argv, const char **envp)
{
  _main(argc, argv, envp);
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x7B7CC140;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x2105B8C9;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xFEA4FB66;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7D1AE6C0;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x2775177E;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x925FBD3;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD987DA09;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xFAD38F20;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xFD4AEBC6;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4C78602E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x9C43184A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xD65BC5BB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3EF2C893;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1B4F0892;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x9CC2356;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7F0EF800;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x6D84E09A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x59CE1C06;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x106AD753;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x879B9FFA;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x6A7D6C68;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x6A458E99;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x6A39546D;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xD3EDBF64;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x78DD036B;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xDDAE7767;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xAAA50E5C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x70AEAEB2;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3102AE18;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x3CECC69A;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA1583D2E;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xCEF3DC6D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3594C51;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xABDC6667;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x456C5FF2;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x6CB61C56;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3A7F577A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x57368E77;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x9F922D9B;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x12ECF301;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x25869D78;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5D870F1A;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xB88B07B2;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x8F8DBC4;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD255375;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7ACF76CF;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBB94A688;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1E6BF741;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD5623A3A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x44EBD4B4;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBBB704D2;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1A3FE1D8;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xE6D571EB;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xF798AB91;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xEE52C1DE;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x550DED4D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x4A7E0D53;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x6F4F1020;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x9DD3B13E;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4C29CAB4;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x16ADA2C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x581B7FEB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xAD7DBBDB;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x50F5A3BE;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBEE0CF2C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xA69D9701;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x333F88BB;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x6A4C98FE;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x56949F54;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xE699052A;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3B1CF719;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xE891D313;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xE0FB59;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xF4BA724D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3145BA6C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1717D7E1;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x3D612168;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4FE65709;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xEA376BDB;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB7039F36;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xAB8C9607;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xE932132E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x6DF54EC1;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xCAC96DBD;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA7FBC65A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7B230A96;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x823EA159;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x6D31027D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xE1D299A8;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xC8814791;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x78F200FC;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x8A561714;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x638F3CA3;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x3C6AFE1E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xFED9B68D;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x9CE3799E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD0B232C8;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4CD3EF18;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA644B0FF;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x2957FEAD;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x48A3CF3B;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5C61789F;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x5868362B;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x2135F32E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xDB14088C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xF2AA0A21;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x4FEB0573;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1B0AA45C;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x7A91AB42;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x78BCF4C8;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD3A6CC4E;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xEA44D817;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x767F67C3;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xC76E48E3;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x8CC053ED;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4C2315F4;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x281C994F;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x11077D71;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x18BBA510;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1B5530B5;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x93828D94;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5F95EEB0;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x2D913F0F;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB76F9152;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xC274D224;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xDA74FED7;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xF448F931;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xEA75D9C9;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x57FFA5B4;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4F3B4B3B;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x871B3FC8;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7725889;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x29C80A0C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x78E70B0C;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x1B57E9B7;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB81BC454;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xE93935C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xC28170AD;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x74A1E673;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x3A6EA904;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x98814E3C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xCF675AFE;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xF4FE401;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xCFD2CE19;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x97A0617;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x70D577BE;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x30458D1C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x59607271;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBD30141F;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x37DFBDD9;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x81C48D18;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xABB4F09E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xEBCB14EC;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x744CD637;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x42A10E5A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7D750737;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xE705255;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x884CB2AB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xCA8E63B7;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xBC827AC1;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xFC556F35;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x38137C4C;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x59D7E723;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB4AB0854;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x67A2E134;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xF0462B20;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xDFD77342;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5DCDAED0;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xC25DD48F;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x76ABE9EE;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x5F28D944;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB9AEC4AB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x7EC18FB1;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xAAF4E86D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x12FE6BAA;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB0EF7F44;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x1174ACFD;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x3780050C;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xB68AF075;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x83A7DC63;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xB56FF91C;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4958B5FD;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x5A51B2CE;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x359AC868;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x5A671B0E;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4D2E377D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x7DA8EDFC;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7C46A290;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBDF48D46;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xD8A6E9F7;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA7EB8AEF;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x47E4EDFB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x86EFCCF8;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x61ECDAE8;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x682795B1;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x365BB07E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x62731532;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1A04803D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD5A5E583;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xE9B10F05;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xDA683933;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1BB5181;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x240D3EA4;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x492AB8FB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x4CAFA828;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xCFB38FD9;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x364D6FA2;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xA4367616;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x1244F9C7;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB502CACA;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xCD123E96;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5CC4BB9A;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD3BFF3E3;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5D0069CB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x1D84208A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xBEA838D2;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xD7D9FC3;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x88339DAA;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x18435C21;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x8A51F2C8;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xC0BB0D11;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x3E9E2B5D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBD00703F;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB614A05C;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x9D376AA8;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x714906EB;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x622C818;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4EDD979E;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA01C1DDE;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x7B98B9AA;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x6E70EE28;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x5D982AF0;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x4D937C7A;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x17DE064C;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x8C843EFF;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x4F3A8634;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xB35FC006;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xE97C7ED9;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x997CB12B;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB0999D5D;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA6BC686;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xF2AB81F5;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xBE0D1038;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xC7B6F558;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x572933A1;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x553EEA26;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x564AAE33;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x50D5CFFA;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xA2CBAF92;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xF702C7A4;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xB2153A9F;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x9D5EC87;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0x570E2F95;
  a1 += (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0x1979F9E8;
  a20 = 4;
  a21 = 5;
  a0 += (a1 + ((a1 >> 5) ^ (16 * a1))) ^ 0xC9F97045;
  a2 = 16 * a0;
  a3 = a0 >> 5;
  a5 = 0xB369A136;
  temp = (a0 + ((a0 >> 5) ^ (16 * a0))) ^ 0xB369A136;
  a1 += temp;
  a30 = 4188039363;
  a31 = 3587894964;
  a0 ^= 4188039363u;
  a1 ^= 3587894964u;
  return 0;
}

六:实现自动化

首先我们分析程序的加密逻辑,知道是一个魔改的TEA

 

编写一个解密函数:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
def decrypt(self):                         # 解密
        self.a1 -= (self.a0 + ((self.a0 >> 5) ^ (16 * self.a0))) ^ self.key_table[self.i + 1]
        self.a1 &= 0xffffffff
        self.a0 -= (self.a1 + ((self.a1 >> 5) ^ (16 * self.a1))) ^ self.key_table[self.i]
        self.a0 &= 0xffffffff
        self.i -= 2
        if self.i == -2:
            a0_string = binascii.unhexlify(hex(self.a0 & 0xffffffff)[2:]).decode("utf-8")
            a0_string = a0_string[::-1]
            a1_string = binascii.unhexlify(hex(self.a1 & 0xffffffff)[2:]).decode("utf-8")
            a1_string = a1_string[::-1]
            self.string = a0_string + a1_string
            print(self.string, end='')
            return
        self.decrypt()

分析下程序的规律

 

我们需要提取的,就是vm指令对应的opcode以及整个字节码

 

图片描述
分析elf的内存,偏移是固定的

 

直接每次读取即可

 

最终脚本:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
# _*_ coding: utf-8 _*_
# editor: SYJ
# function: Reversed By SYJ
"""
describe:
 
"""
from pwn import *
import hashlib
 
class Decompile:
    def __init__(self, file_name, opcode_keys_, code):
        self.filename = file_name   # 要逆向的文件路径
        self.asm_code = []          # 反汇编的指令列表
        self.pc = 0                 # 虚拟机的rip
        self.code = code            # 存放文件的虚拟code
        self.opcode_keys = opcode_keys_  # 存放文件的指令集代表的opcode
        self.recompile_to_c = ""    # 二维数组来存放每个文件反编译成c的代码
        self.key_table = []         # 二维数组来存放文件的加密用的key_table
        self.enc_number = 0         # 根据key来判断加密的次数
        self.a0 = 0                 # 加密之后的数据(后面得到key_table之后会设置)
        self.a1 = 0
        self.i = 0                 # 解密的递归循环变量,用来判断是否退出递归
        self.string = ''           # 解密之后的字符串
 
    def recompile(self):      # 定义一个方法来重编译(没必要,指令很简单,没有涉及栈操作)
        memory = bytes(self.code)
        data_bin = memory      # 前面划分一段区域来当内存,直接将我们的opcode存放进去
        asm_code = "\n".join(self.asm_code)
        code_bin = asm(asm_code, arch="i386")
        file = "{}_bin".format(string[len(self.filename)-10:len(self.filename)-4])
        open(file, "wb").write(data_bin + code_bin)
 
    def get_opcode(self):          # 获取opcode
        opcode = self.code[self.pc]
        return opcode
 
    def rip_ad_three(self):       # 获取rip跳转和操作数
        fuck_num1 = self.code[self.pc+1]
        fuck_num2 = self.code[self.pc+2]
        self.pc += 3
        return fuck_num1, fuck_num2
 
    def rip_ad_five(self):
        fuck_num1 = self.code[self.pc+1]
        self.pc += 6
        return fuck_num1
 
    def used_by_jcc(self):          # 根据虚拟机实现跳转指令的方式(分析出来获取比较值和跳转值)
        fuck_num1 = self.code[self.pc + 1]
        self.pc += 5
        val = self.code[self.pc]
        self.pc += 1
        return fuck_num1, val
 
    def rip_ad_six(self):            # 这个里面有点特殊,除开获取跳转和操作数之外,还要获取key以及key_len
        fuck_num1 = self.code[self.pc+1]
        fuck_num2 = bytes(self.code[self.pc+2:self.pc+6])[::-1]
        hex_val = '0x'               # 该值用来返回
        for x in fuck_num2:
            hex_val += "{:02x}".format(x)
        self.pc += 6
        if hex_val != '0x00000004' and hex_val != '0x00000005':
            self.key_table.append(eval(hex_val))         # eval自动识别
        return fuck_num1, hex_val
 
    def disamone(self):          # 反编译为伪汇编指令
        regs = ['eax', 'ebx', 'ecx', 'edx', 'esi', 'edi', 'reg6', 'reg7', 'reg8', 'reg9', 'reg10', 'reg11', 'reg12', 'reg12', 'reg14', 'reg15', 'reg16', 'reg17', 'reg18', 'reg19', 'eax', 'ebx', 'reg22', 'reg23', 'reg24', 'reg25', 'reg26', 'reg27', 'reg28', 'reg29', 'ecx', 'edx', 'reg32']
        opcode = self.get_opcode()
        offset = len(self.code)
        if opcode == self.opcode_keys[11]:
            pc = self.pc
            num1,num2 = self.rip_ad_six()
            self.asm_code.append("_%s: mov %s, %s" % (hex(offset+pc), regs[num1], num2))
        elif opcode == self.opcode_keys[0]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov %s, %s" % (hex(offset+pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[1]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: add %s, %s" % (hex(offset+pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[2]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: shl ecx, al" % (hex(offset + pc)))        # regs[num2] % (hex(offset + pc), regs[num1])
        elif opcode == self.opcode_keys[3]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: shr edx, bl" % (hex(offset + pc)))       # , regs[num2] % (hex(offset + pc), regs[num1])
        elif opcode == self.opcode_keys[4]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov eax, %s\ncdq\nidiv %s\nmov %s, edx" % (hex(offset + pc), regs[num1], regs[num2], regs[num1]))
        elif opcode == self.opcode_keys[7]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: xor %s, %s" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[5]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov %s, [%s+0x86]" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[6]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: mov [%s+0x86], %s" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[8]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.asm_code.append("_%s: cmp %s, %s" % (hex(offset + pc), regs[num1], regs[num2]))
        elif opcode == self.opcode_keys[9]:
            pc = self.pc
            num1 = self.rip_ad_five()
            self.asm_code.append("_%s: nop" % (hex(offset + pc)))
        elif opcode == self.opcode_keys[10]:
            pc = self.pc
            num1, val = self.used_by_jcc()
            jcc = ['jmp', 'jl', 'jg']
            temp = ''
            if val == 0:
                temp = jcc[0]
            elif val == 1:
                temp = jcc[1]
            elif val == 2:
                temp = jcc[2]
            self.asm_code.append("_%s: %s %s" % (hex(offset + pc), temp, num1))
        else:
            print("undefined instruction")
 
    def disamtoc(self):                # 根据分析反编为C代码
        regs = ['a0', 'a1', 'a2', 'a3', 'a4', 'a5', 'a6', 'a7', 'a8', 'a9', 'a10', 'a11', 'a12', 'a13', 'a14', 'a15', 'a16', 'a17', 'a18', 'a19', 'a20', 'a21', 'a22', 'a23', 'a24', 'a25', 'a26', 'a27', 'a28', 'a29', 'a30', 'a31', 'a32']
        opcode = self.get_opcode()
        if opcode == self.opcode_keys[11]:
            pc = self.pc
            num1, num2 = self.rip_ad_six()
            # self.recompile_to_c += "_{}: {}={};\n".format(hex(pc), regs[num1], num2)
            self.recompile_to_c += "{}={};\n".format(regs[num1], num2)
        elif opcode == self.opcode_keys[0]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}={};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[1]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}+{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}+={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[2]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}<<{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}<<={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[3]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}>>{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}>>={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[4]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {} % {};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{} %= {};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[7]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}^{};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{} ^= {};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[5]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: {}=arr[{}+0x86];\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "{}=arr[{}+0x86];\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[6]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            # self.recompile_to_c += "_{}: arr[{}+0x86]={};\n".format(hex(pc), regs[num1], regs[num2])
            self.recompile_to_c += "arr[{}+0x86]={};\n".format(regs[num1], regs[num2])
        elif opcode == self.opcode_keys[8]:
            pc = self.pc
            num1, num2 = self.rip_ad_three()
            self.recompile_to_c += "eflag={}-{};\n".format(regs[num1], regs[num2])
            # self.recompile_to_c += "_{}: eflag={}-{};\n".format(hex(pc), regs[num1], regs[num2])
        elif opcode == self.opcode_keys[9]:
            pc = self.pc
            num1 = self.rip_ad_five()
            # self.recompile_to_c += "_{}: goto _{}".format(hex(pc), hex(pc+1))
            self.recompile_to_c += "\n"
        elif opcode == self.opcode_keys[10]:
            pc = self.pc
            num1, val = self.used_by_jcc()
            jcc = ['jmp', 'jl', 'jg']
            temp = ''
            if val == 0:
                temp = jcc[0]
            elif val == 1:
                temp = jcc[1]
            elif val == 2:
                temp = jcc[2]
            self.recompile_to_c += "\n"
            # self.recompile_to_c += "_{}: val={};\nif(val == 0):goto _{};\nif(val==1 && eflag<0):goto _{}\nif(val==2 && eflag>0):goto _{}\n".format(hex(pc), val, hex(pc+num1), hex(pc+num1), hex(pc+num1))
        else:
            print("undefined instruction")
 
    def get_asm_code(self):         # 获取伪汇编代码
        while self.pc < len(self.code):
            self.disamone()
        print("\n".join(self.asm_code))
 
    def get_c_code_and_encinfo(self):           # 获取重编的C代码和加密的一些信息
        while self.pc < len(self.code):
            self.disamtoc()
        self.a0 = self.key_table[len(self.key_table)-2]
        self.a1 = self.key_table[len(self.key_table)-1]
        self.enc_number = (len(self.key_table) - 2)
        self.i = len(self.key_table) - 4
        # print(self.i)
        # print(self.enc_number)      # 打印key的数量
        # print(self.recompile_to_c)  # print c_code
        # print(self.key_table)       # 打印key_table
        # print(self.a0)
        # print(self.a1)
 
    def decrypt(self):                         # 解密
        self.a1 -= (self.a0 + ((self.a0 >> 5) ^ (16 * self.a0))) ^ self.key_table[self.i + 1]
        self.a1 &= 0xffffffff
        self.a0 -= (self.a1 + ((self.a1 >> 5) ^ (16 * self.a1))) ^ self.key_table[self.i]
        self.a0 &= 0xffffffff
        self.i -= 2
        if self.i == -2:
            a0_string = binascii.unhexlify(hex(self.a0 & 0xffffffff)[2:]).decode("utf-8")
            a0_string = a0_string[::-1]
            a1_string = binascii.unhexlify(hex(self.a1 & 0xffffffff)[2:]).decode("utf-8")
            a1_string = a1_string[::-1]
            self.string = a0_string + a1_string
            print(self.string, end='')
            return
        self.decrypt()
 
path = 'D:\\ReverseSource\\2021wmx-VMautomatic'       # 所在的文件夹路径
file_number = 16                                      # 文件数量
for i in range(file_number):
    path += '\\solve{}.exe'.format(i)                 # 文件绝对路径
    data_bin = open(path, 'rb').read()                # 文件的bin数据
 
    opcode_length = data_bin[0xF390: 0xF390+4]        # 根据文件偏移为0x0000000000010390 - 0x1000获取opcode长度
    opcode_length = opcode_length[::-1]
    opcode_length_hex = '0x'
    for x in range(len(opcode_length)):
        opcode_length_hex += "{:02x}".format(opcode_length[x])
    opcode_length_value = eval(opcode_length_hex)
    # print(hex(opcode_length_value))
 
    opcode_keys = []                                  # 根据文件偏移都为0x4020-0x1000获取opcode_keys
    for j in range(12):
        opcode_keys.append(data_bin[0x3020+j])
    # print(opcode_keys)
 
    opcode = data_bin[0x3040: 0x3040+opcode_length_value]      # 根据文件偏移为0x4040-0x1000得出虚拟机的code,读取长度为前面得到的opcode_length_value
    opcode_arr = []
    for z in range(opcode_length_value):
        opcode_arr.append(opcode[z])
    # print(opcode_arr)
 
    temp_object = Decompile(path, opcode_keys, opcode_arr)
    # temp_object.get_asm_code()
    temp_object.get_c_code_and_encinfo()
    temp_object.decrypt()
    path = "D:\\ReverseSource\\2021wmx-VMautomatic"
 
print("\n")
# 将控制台输出的字符串全部md5即可得到flag
all_solve = "ozeletrtlhqtbthxqtkclucqtrkmrpguwewbzocqwnvhyxrxkdgfeqdefoucorbsbidzeijvcncopedpkfbezltqvebskounguxsvojwpgocejgmjdaxfscjflhjpqts"
print('flag{'+ hashlib.md5(all_solve.encode('utf-8')).hexdigest().lower() + '}')

个人notion笔记地址:https://agate-colony-3f5.notion.site/CTF-VM-94d5b358d2b946b0a64f9910030f78af


[招生]科锐逆向工程师培训(2024年11月15日实地,远程教学同时开班, 第51期)

最后于 2021-9-30 03:24 被SYJ-Re编辑 ,原因:
上传的附件:
收藏
免费 8
支持
分享
最新回复 (3)
雪    币: 7325
活跃值: (3803)
能力值: (RANK:1130 )
在线值:
发帖
回帖
粉丝
2
现在的CTF题目,都这么牛逼了?
2021-9-30 11:32
1
雪    币: 515
能力值: ( LV1,RANK:0 )
在线值:
发帖
回帖
粉丝
3

简易版, 不需要pwntools

from ctypes import *
import struct
from hashlib import md5


def load(file_name):
    with open(file_name, 'rb') as file:
        op_dict = {}
        file.seek(0xf390)
        size = file.read(4)
        size = struct.unpack("I", size)[0]
        file.seek(0x3040)
        bytecode = file.read(size)
        file.seek(0x3020)
        op_code = file.read(12)
        op_dict[op_code[0]] = '='
        op_dict[op_code[1]] = '+='
        op_dict[op_code[2]] = '<<='
        op_dict[op_code[3]] = '>>='
        op_dict[op_code[4]] = '%='
        op_dict[op_code[7]] = '^='
        op_dict['jmp'] = op_code[9]
        op_dict['mov'] = op_code[11]
        return bytecode, op_dict


def pack2int(a, b, c, d, type):
    return c_int((d << 24) + (c << 16) + (b << 8) + a).value if type else c_uint(
        (d << 24) + (c << 16) + (b << 8) + a).value


def translate(bytecode, op_dict):
    number_list = []
    bytecode_size = len(bytecode)
    ip = 0
    while ip < bytecode_size:
        # print("{:#x}: ".format(ip), end='')
        op_code = bytecode[ip]
        ip += 1
        if op_code in op_dict:
            # print("reg[{:#x}] {} reg[{:#x}]".format(bytecode[ip], op_dict[op_code], bytecode[ip + 1]))
            ip += 2
        elif op_code == op_dict['jmp']:
            number = pack2int(bytecode[ip], bytecode[ip + 1], bytecode[ip + 2], bytecode[ip + 3], 1)
            ip += 4
            # print("jmp {:#x}".format(ip + number))
            ip = ip + number
        elif op_code == op_dict['mov']:
            number = pack2int(bytecode[ip + 1], bytecode[ip + 2], bytecode[ip + 3], bytecode[ip + 4], 0)
            if number not in [4, 5]:
                number_list.append(number)
            # print("reg[{:#x}] = {:#x}".format(bytecode[ip], number))
            ip += 5
    return number_list


def decrypt(number_list):
    flag_part2 = number_list.pop()
    flag_part1 = number_list.pop()
    for i in range(128):
        key = number_list.pop()
        flag_part2 -= (((flag_part1 << 4) ^ (flag_part1 >> 5)) + flag_part1) ^ key
        flag_part2 &= 0xffffffff
        key = number_list.pop()
        flag_part1 -= (((flag_part2 << 4) ^ (flag_part2 >> 5)) + flag_part2) ^ key
        flag_part1 &= 0xffffffff
    return flag_part1, flag_part2


def main():
    filename = "output/solve/solve{}.exe"
    flag = b''
    for i in range(100):
        bytecode, op_dict = load(filename.format(i))
        number_list = translate(bytecode, op_dict)
        result = decrypt(number_list)
        flag += result[0].to_bytes(4, byteorder='little') + result[1].to_bytes(4, byteorder='little')
    print(flag)
    print(md5(flag).hexdigest())


if __name__ == '__main__':
    main()


最后于 2021-11-6 12:36 被LS_恶霸崔佛编辑 ,原因:
2021-11-6 12:26
2
雪    币: 576
活跃值: (2035)
能力值: ( LV2,RANK:10 )
在线值:
发帖
回帖
粉丝
4

。。范围理解错了.感谢分享...

最后于 2021-11-6 19:08 被kakasasa编辑 ,原因:
2021-11-6 16:34
0
游客
登录 | 注册 方可回帖
返回
//