奇安信攻防社区-恶意样本自动化配置提取初探

恶意样本自动化配置提取初探

本文分享自己对 capev2 上 emotet 配置提取脚本的参考分析和探究历程，希望能给到大家知识帮助。

0x00 前言：
========

本篇参考 github 上 [CAPEv2](https://github.com/kevoreilly/CAPEv2/blob/f2ab891a278b2875c79b4f2916d086f870b54ed5/modules/processing/parsers/CAPE/Emotet.py) 沙箱的提取代码，在前面[奇安信攻防社区-APT 恶意 DLL 分析及 C2 配置提取（子 DLL 篇） (butian.net)](https://forum.butian.net/share/1804)分析的基础上尝试编写自动化配置提取，如有错误还请指正。

0x01 编写环境：
==========

**语言：** python

**外部库：**

yara——匹配规则，锁定 C2 配置及密钥配置位置，pip install yara-python

Cryptodome——提取整合加密密钥并导出 pip install pycryptodomex

pefile——应用 PE 结构模板，定位文件头和节表区的字段和数据

**标准库：**  
struct、socket、itertools

**编译器**：

vscode

0x02 样本IOC：
===========

| HASH | 值 |
|---|---|
| MD5 | 4e22717b48f2f75fcfd47531c780b218 |
| SHA1 | 60b637e95b1f2d14faaa71085b7e26321bfeeb6d |
| SHA256 | 7f94107c9becbcc6ca42070fca7e1e63f29cdd85cbbd8953bbca32a1b4f91219 |

0x03 ECC 密钥提取：
==============

首先在 IDA 或 XDBG 中定位到解密的代码的特征数据区，由于之前分析得很详细了，所以我们直接放上截图：（蓝框就是我们认定的特征区）

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-c72997b04a250844dacb2f8b239831d1c3ada01b.png)

编写密钥区的 Yara 规则：
---------------

**Yara 规则怎么写呢，把地址部分的都模糊查询，指令码部分的字节都一一对应：**

比如说上面蓝框的第一行 FF B4 24 A8 01 00 00 push \[esp+28Ch+var\_E4\] ，由于 IDA 中在识别函数的过程中插入了 var\_E4 变量，所以我们对确切的地址部分有点模糊，我们看 xdbg 中的同样位置。

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-6f6ef7124727eef103642624be9d1a502a77ec33.png)

可以发现真实的指令应该是 pust dword ptr ss:\[esp+1A8\]，根据 x86 指令码和机器码的转换大概可以确定 FF B4 对应着 push，24 对应着后面的 esp 的基地偏移量寻址，所以第一行我们提取出的 Yara 规则是 FF B4 \[3\] 00 00。

最后两个全 0 字节是因为这里是基于 ss 16 位段选择为基址的，而程序是 32 位，所以开头的 2 个字节 16 位就一定会空下来。

其它行同理，所以最后蓝框中的 Yara 规则就是 *{FF B4 \[3\] 00 00 FF B4 \[3\] 00 00 8B 94 \[3\] 00 00 E8 \[4\] 83 C4 0C 89 84 \[3\] 00 00 8D 84 \[3\] 00 00 B9 \[4\] 50 FF B4 \[3\] 00 00 FF B4 \[3\] 00 00 8B 94 \[3\] 00 00 E8}*。

**写成 Yara 规则代码就是：**

```python
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet ECC Extra"
    strings:
        $ref_ecc = {FF B4 [3] 00 00 FF B4 [3] 00 00 8B 94 [3] 00 00 E8 [4] 83 C4 0C 89 84 [3] 00 00 8D 84 [3] 00 00 B9 [4] 50 FF B4 [3]00 00 FF B4 [3]00 00 8B 94 [3]00 00 E8}
    condition:
        $ref_ecc            
}
"""
```

利用 Yara 库 API 定位特征区首地址：
-----------------------

首先参考官方文档了解 Yara API 和 类对象：[在 Python 中使用 YARA — yara 4.2.0 文档](https://yara.readthedocs.io/en/v4.2.3/yarapython.html)

**我们要用到的 API 如下：**

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-14537b7616434e659d8353eb257bf732eec943c7.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-15fbefd899de839bdfd757b5b1cab25a75e02c6f.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-f9683fbc036649040422f2ec2ba9bc2ea76393e1.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-fc7d56bde4e67149d551f789eaad01d6224a6cb1.png)

**懂了之后就尝试编写代码获取特征区首地址了：**

我们这里用的 yara 是基于静态扫描，也就是说它不会展开内存来匹配，所以匹配都是基于文件字节码的，返回值也是特征区在文件中的匹配，特别要注意的是它返回的偏移是 10 进制的，所以我们要自己转为十六进制。

```python
import yara
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet ECC Extra"
    strings:
        $ref_ecc = {FF B4 [3] 00 00 FF B4 [3] 00 00 8B 94 [3] 00 00 E8 [4] 83 C4 0C 89 84 [3] 00 00 8D 84 [3] 00 00 B9 [4] 50 FF B4 [3]00 00 FF B4 [3]00 00 8B 94 [3]00 00 E8}
    condition:
        $ref_ecc            
}
"""
def yara_scan(raw_data):
    addresses = {}
    yara_rules = yara.compile(source=rule_source)
    matches = yara_rules.match(data=raw_data)
    for match in matches:
        for item in match.strings:
            addresses[item[1]] = hex(item[0])       #转为 16 进制，方便查看
    return addresses

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(yara_scan(file_data))                     #返回的结果为 {'$ref_ecc': '0xee6d'}
```

**验证一下：**  
![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-ac899d9b22c69fbb2a592fa31136455c1719e415.png)

从特征区首中定位要解密的数据区：
----------------

两个数据分别在 0xee6d 起始的特征区中 -5 和 +44 处，我们可以设两个变量为 delta1 = -5 , delta2 = 44; 后面写代码时会用到。

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-b55f6128b4b021157ea18593258f697e6dfee1cf.png)

因为在编译时变量引用都被替换编译成 VA 了，所以我们需要把 VA 转 RVA，再获取 FOA（文件偏移），这需要另一个外部库 [pefile — pefile documentation](https://pefile.readthedocs.io/en/latest/modules/pefile.html)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-5034470d841ef979d42d531cec7eaf953b3fd4db.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-c7785d0b784cafab89d543d762887829a725af1f.png)

**编写脚本如下：**

```python
import yara
import pefile
import struct
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet ECC Extra"
    strings:
        $ref_ecc = {FF B4 [3] 00 00 FF B4 [3] 00 00 8B 94 [3] 00 00 E8 [4] 83 C4 0C 89 84 [3] 00 00 8D 84 [3] 00 00 B9 [4] 50 FF B4 [3]00 00 FF B4 [3]00 00 8B 94 [3]00 00 E8}
    condition:
        $ref_ecc            
}
"""
def yara_scan(raw_data):
    addresses = {}
    yara_rules = yara.compile(source=rule_source)
    matches = yara_rules.match(data=raw_data)
    for match in matches:
        for item in match.strings:
            addresses[item[1]] = hex(item[0])       #手动转为 16 进制，方便查看
    return addresses

def positioning_data(filebuf):
    conf_dict = {}
    pe = None
    pe = pefile.PE(data=filebuf, fast_load=False)
    image_base = pe.OPTIONAL_HEADER.ImageBase       #获取载入基址，用于从 VA 转 RVA
    yara_matches = yara_scan(filebuf)
    if yara_matches.get("$ref_ecc"):
        ref_ecc_offset = int(yara_matches["$ref_ecc"],16)
        delta1 = -5
        delta2 = 44
        ref_eck_rva = struct.unpack("I", filebuf[ref_ecc_offset + delta1 : ref_ecc_offset + delta1 + 4])[0] - image_base    #struct.unpack(format, buffer)，根据格式字符串 format 从缓冲区 buffer 解包，返回元祖，所以这里用[0]来提取。I 是 unsigned int 类型
        ref_ecs_rva = struct.unpack("I", filebuf[ref_ecc_offset + delta2 : ref_ecc_offset + delta2 + 4])[0] - image_base    
        eck_offset = pe.get_offset_from_rva(ref_eck_rva)        #获取此 RVA 对应的文件偏移量。
        ecs_offset = pe.get_offset_from_rva(ref_ecs_rva)
        return "eck_offset:",hex(eck_offset),"ecs_offset",hex(ecs_offset)

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(positioning_data(file_data))                      #返回的结果为('eck_offset:', '0xb34', 'ecs_offset', '0xb94')
```

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-6d6cd19d681a9800b8c3e8400d6c4f18f04f822b.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-b8eb905758867a464fbf912af17b22ab2c8cdb86.png)

编写 ECC 解密代码：（成功提取）
------------------

以前的分析中说过了公钥在加密中的数据格式，第一个 Dword 是解密的 key，第二个 Dword 是公钥的长度，剩下的是加密的数据。

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-45ad1559ac6211003781a7fa3f9cc26a99ec13f7.png)

我们可以用一个 xor 函数来实现异或解密并依旧用 struct.unpack 来把 4 字节格式化输出，因为一个 key 要重复对后面的数据解密使用，所以我们这里用 python 标准库 itertools 的 API [itertools --- 为高效循环而创建迭代器的函数 — Python 3.10.6 文档](https://docs.python.org/zh-cn/3/library/itertools.html?highlight=cycle#itertools.cycle)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-dab18f3b0dc14c6d6e441cd1bfd1bfdd94223162.png)

**xor 函数如下：**

```python
from itertools import cycle
def xor_data(data, key):
    return bytes(c ^ k for c, k in zip(data, cycle(key)))       
#将可迭代的对象作为参数，将对象中对应的元素打包成一个个元组，然后返回由这些元组组成的列表。cycle不断返回一样的副本。
#所以返回类似于[(data1,key),(data2,key)……],然后用列表推导式从中获取元祖的两个元素
```

但是输出的是 ECC 密钥并不是可读的，因为它们只是一串字节码，我们需要把它格式化为 ECC 密钥该有的形式，可以从 pycryptodome 官方文档中找到可用的 API [ECC — PyCryptodome 3.15.0 文档](https://pycryptodome.readthedocs.io/en/latest/src/public_key/ecc.html#ecc-table)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-5371c8f146ef274fdd722607888cf792e9773e38.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-d26f2e241b279c075233b382afff2569391412b6.png)

**所以最终的 ECC 密钥提取脚本如下：**

```python
import yara
import pefile
import struct
from Cryptodome.PublicKey import ECC
from itertools import cycle
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet ECC Extra"
    strings:
        $ref_ecc = {FF B4 [3] 00 00 FF B4 [3] 00 00 8B 94 [3] 00 00 E8 [4] 83 C4 0C 89 84 [3] 00 00 8D 84 [3] 00 00 B9 [4] 50 FF B4 [3]00 00 FF B4 [3]00 00 8B 94 [3]00 00 E8}
    condition:
        $ref_ecc            
}
"""
def yara_scan(raw_data):
    addresses = {}
    yara_rules = yara.compile(source=rule_source)
    matches = yara_rules.match(data=raw_data)
    for match in matches:
        for item in match.strings:
            addresses[item[1]] = hex(item[0])       #手动转为 16 进制，方便查看
    return addresses

def xor_data(data, key):
    return bytes(c ^ k for c, k in zip(data, cycle(key)))       
#将可迭代的对象作为参数，将对象中对应的元素打包成一个个元组，然后返回由这些元组组成的列表。cycle不断返回一样的副本。
#所以返回类似于[(data1,key),(data2,key)……],然后用列表推导式从中获取元祖的两个元素

def extract_ecc(filebuf):
    conf_dict = {}
    pe = None
    pe = pefile.PE(data=filebuf, fast_load=False)
    image_base = pe.OPTIONAL_HEADER.ImageBase       #获取载入基址，用于从 VA 转 RVA
    yara_matches = yara_scan(filebuf)
    if yara_matches.get("$ref_ecc"):
        ref_ecc_offset = int(yara_matches["$ref_ecc"],16)
        delta1 = -5
        delta2 = 44
        ref_eck_rva = struct.unpack("I", filebuf[ref_ecc_offset + delta1 : ref_ecc_offset + delta1 + 4])[0] - image_base    #struct.unpack(format, buffer)，根据格式字符串 format 从缓冲区 buffer 解包，返回元祖，所以这里用[0]来提取。
        ref_ecs_rva = struct.unpack("I", filebuf[ref_ecc_offset + delta2 : ref_ecc_offset + delta2 + 4])[0] - image_base    #struct.unpack(format, buffer)，根据格式字符串 format 从缓冲区 buffer 解包，返回元祖，所以这里用[0]来提取。
        eck_offset = pe.get_offset_from_rva(ref_eck_rva)
        ecs_offset = pe.get_offset_from_rva(ref_ecs_rva)

key = filebuf[eck_offset : eck_offset + 4]
        size = struct.unpack("I", filebuf[eck_offset + 4 : eck_offset + 8])[0] ^ struct.unpack("I", key)[0]
        eck_offset += 8
        eck_key = xor_data(filebuf[eck_offset : eck_offset + size], key)
        key_len = struct.unpack("<I", eck_key[4:8])[0]          #ECC密钥还有长度的？
        conf_dict.setdefault(
            "ECC ECK1",
            ECC.construct(
                curve="p256",
                point_x=int.from_bytes(eck_key[8 : 8 + key_len], "big"),
                point_y=int.from_bytes(eck_key[8 + key_len :], "big"),
            ).export_key(format="PEM"),
        )

key = filebuf[ecs_offset : ecs_offset + 4]
        size = struct.unpack("I", filebuf[ecs_offset + 4 : ecs_offset + 8])[0] ^ struct.unpack("I", key)[0]
        ecs_offset += 8
        ecs_key = xor_data(filebuf[ecs_offset : ecs_offset + size], key)
        key_len = struct.unpack("<I", ecs_key[4:8])[0]
        conf_dict.setdefault(
            "ECC ECS1",
            ECC.construct(
                curve="p256",
                point_x=int.from_bytes(ecs_key[8 : 8 + key_len], "big"),
                point_y=int.from_bytes(ecs_key[8 + key_len :], "big"),
            ).export_key(format="PEM"),
        )
    return conf_dict

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(extract_ecc(file_data))       #最终输出{'ECC ECK1': '-----BEGIN PUBLIC KEY-----\nMFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAE86M1tQ4uK/Q1Vs0KTCk+fPEQ3cuw\nTyCz+gIgzky2DB5Elr60DubJW5q9Tr2dj8/gEFs0TIIEJgLTuqzx+58sdg==\n-----END PUBLIC KEY-----', 'ECC ECS1': '-----BEGIN PUBLIC KEY-----\nMFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAEQF90tsTY3Aw9HwZ6N9y5+be9Xoov\npqHyD6F5DRTl9THosAoePIs/e5AdJiYxhmV8Gq3Zw1ysSPBghxjZdDxY+Q==\n-----END PUBLIC KEY-----'}
```

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-777a12366137679d2dd9c28795704280aae58589.png)

0x04 C2 配置提取：
=============

还是一样先定位到特征数据区，由于用的同一个解密函数，所以我们可以直接用 IDA 的热键 X 来交叉引用来寻找第三个，也就是 C2 配置区。

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-25d9cc2e15a4411cd989e18602e589dd84f54120.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-4d2f54d5455d0b2e939d4ef3714d28c8d19376d7.png)

同理编写 C2 区的 Yara 规则：
-------------------

同理，参考前面的密钥区的 Yara 规则，地址部分的都模糊查询，指令码部分的字节都一一对应。所以蓝框区的 Yara 规则就是 *{FF 74 \[2\] FF 74 \[2\] 8B 54 \[2\] E8 \[4\] 8B 54 \[2\] 83 C4 0C 89 44 \[2\] 8B F8 03 44 \[2\] B9 \[4\] 89 44 \[2\] E9}*

**合成 Yara 规则代码就是：**

```python
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet C2 Extra"
    strings:
        $snippet = {FF 74 [2] FF 74 [2] 8B 54 [2] E8 [4] 8B 54 [2] 83 C4 0C 89 44 [2] 8B F8 03 44 [2] B9 [4] 89 44 [2] E9}
    condition:
        $snippet        
}
"""
```

同理定位特征区首地址：
-----------

```python
import yara
import pefile
import struct
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet C2 Extra"
    strings:
        $ref_c2 = {FF 74 [2] FF 74 [2] 8B 54 [2] E8 [4] 8B 54 [2] 83 C4 0C 89 44 [2] 8B F8 03 44 [2] B9 [4] 89 44 [2] E9}
    condition:
        $ref_c2     
}
"""
def yara_scan2(raw_data):
    addresses = {}
    yara_rules = yara.compile(source=rule_source)
    matches = yara_rules.match(data=raw_data)
    for match in matches:
        for item in match.strings:
            addresses[item[1]] = hex(item[0])       #手动转为 16 进制，方便查看
    return addresses

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(yara_scan2(file_data))                        #返回的结果为{'$ref_c2': '0x12ca2'}
```

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-7a56268f84a054058414e09454e307689def54cd.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-571a3b3c5f960381bf0caeb698794067cc909805.png)

同理定位要解密的数据区:
------------

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-5be6aac8db138d28ce833f77e47a7885375136a5.png)

**编写脚本如下：**

def positioning_c2_data(filebuf):
    conf_dict = {}
    pe = None
    pe = pefile.PE(data=filebuf, fast_load=False)
    image_base = pe.OPTIONAL_HEADER.ImageBase       #获取载入基址，用于从 VA 转 RVA
    yara_matches = yara_scan2(filebuf)
    if yara_matches.get("$ref_c2"):
        delta = -5
        c2list_va_offset = int(yara_matches["$ref_c2"],16)
        c2_list_va = struct.unpack("I", filebuf[c2list_va_offset + delta : c2list_va_offset + delta + 4])[0]
        c2_list_rva = c2_list_va - image_base
        c2_list_offset = pe.get_offset_from_rva(c2_list_rva)
    return "c2_list_offset:",hex(c2_list_offset)

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(positioning_c2_data(file_data))                       #返回的结果为('c2_list_offset:', '0x21e00')
```

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-d8510d6da76fbb4a4316e1b9e941d1baee7df9e4.png)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-c19e2a2d2fbe871870d0ae1a71b277fd2c18b00f.png)

同理编写 C2 解密代码：（成功提取）
-------------------

以前的分析中说过了公钥在加密中的数据格式，第一个 Dword 是解密的 key，第二个 Dword 是公钥的长度，剩下的是加密的数据，其中 C2 数据格式如下，以 8 个字节为一个单位。

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-bf0a182a54deccfd91c50c3a506c41a174480912.png)

和前面一样，我们使用 xor 函数解密，不同的是这里提取的是 IP ，所以我们需要引用 IP 相关的标准库 socket。对于 IP 数据我们先用 struct.unpack 把 4 字节区域格式化整合出来，再用 socket 库的 inet\_ntoa API 把其转换成点分十进制形式。对于端口数据我们直接用 struct.unpack 把 2 字节区域整合出来即可。[socket --- 底层网络接口 — Python 3.10.6 文档](https://docs.python.org/zh-cn/3/library/socket.html?highlight=socket#module-socket)

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-c82d24bdaebe2f640656f80401c1a128f246becc.png)

**所以最终的 C2 密钥提取脚本如下：**

```python
import yara
import pefile
import struct
from itertools import cycle
import socket
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet C2 Extra"
    strings:
        $ref_c2 = {FF 74 [2] FF 74 [2] 8B 54 [2] E8 [4] 8B 54 [2] 83 C4 0C 89 44 [2] 8B F8 03 44 [2] B9 [4] 89 44 [2] E9}
    condition:
        $ref_c2     
}
"""
def yara_scan2(raw_data):
    addresses = {}
    yara_rules = yara.compile(source=rule_source)
    matches = yara_rules.match(data=raw_data)
    for match in matches:
        for item in match.strings:
            addresses[item[1]] = hex(item[0])       #手动转为 16 进制，方便查看
    return addresses

def xor_data(data, key):
    return bytes(c ^ k for c, k in zip(data, cycle(key)))

def extra_c2_data(filebuf):
    conf_dict = {}
    pe = None
    pe = pefile.PE(data=filebuf, fast_load=False)
    image_base = pe.OPTIONAL_HEADER.ImageBase       #获取载入基址，用于从 VA 转 RVA
    yara_matches = yara_scan2(filebuf)
    if yara_matches.get("$ref_c2"):
        delta = -5
        c2list_va_offset = int(yara_matches["$ref_c2"],16)
        c2_list_va = struct.unpack("I", filebuf[c2list_va_offset + delta : c2list_va_offset + delta + 4])[0]
        c2_list_rva = c2_list_va - image_base
        c2_list_offset = pe.get_offset_from_rva(c2_list_rva)
        key = filebuf[c2_list_offset : c2_list_offset + 4]
        presize = filebuf[c2_list_offset + 4 : c2_list_offset + 8]
        size = struct.unpack("I", presize)[0] ^ struct.unpack("I", key)[0]
        c2_list_offset += 8
        c2_list = xor_data(filebuf[c2_list_offset:], key)
        offset = 0
        while offset < size:
            ip = struct.unpack(">I", c2_list[offset : offset + 4])[0]
            c2_address = socket.inet_ntoa(struct.pack("!L", ip))        #将 32 位压缩 IPv4 地址（一个 类字节对象，长 4 个字节）转换为标准的点分十进制字符串形式（如 '123.45.67.89' ）
            port = str(struct.unpack(">H", c2_list[offset + 4 : offset + 6])[0])
            if not c2_address or not port:
                break
            conf_dict.setdefault("address", []).append(f"{c2_address}:{port}")
            c2found = True
            offset += 8    
    return conf_dict

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(extra_c2_data(file_data))                     #返回的结果为{'address': ['131.100.24.231:80', '209.59.138.75:7080', '103.8.26.103:8080', '51.38.71.0:443', '212.237.17.99:8080', '79.172.212.216:8080', '207.38.84.195:8080', '104.168.155.129:8080', '178.79.147.66:8080', '46.55.222.11:443', '103.8.26.102:8080', '192.254.71.210:443', '45.176.232.124:443', '203.114.109.124:443', '51.68.175.8:8080', '58.227.42.236:80', '45.142.114.231:8080', '217.182.143.207:443', '178.63.25.185:443', '45.118.115.99:8080', '103.75.201.2:443', '104.251.214.46:8080', '158.69.222.101:443', '81.0.236.90:443', '45.118.135.203:7080', '176.104.106.96:8080', '212.237.56.116:7080', '216.158.226.206:443', '173.212.193.249:8080', '50.116.54.215:443', '138.185.72.26:8080', '41.76.108.46:8080', '212.237.5.209:443', '107.182.225.142:8080', '195.154.133.20:443', '162.214.50.39:7080', '110.232.117.186:8080']}
```

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-a22b7bc8f20f6329ec35ab058d24477d37cdc8d5.png)

0x05 整合代码如下：
============

```python
import yara
import pefile
import struct
from Cryptodome.PublicKey import ECC
from itertools import cycle
import socket
rule_source = """
rule Emotet
{
    meta:
        description = "Emotet ECC Extra"
    strings:
        $ref_c2 = {FF 74 [2] FF 74 [2] 8B 54 [2] E8 [4] 8B 54 [2] 83 C4 0C 89 44 [2] 8B F8 03 44 [2] B9 [4] 89 44 [2] E9}
        $ref_ecc = {FF B4 [3] 00 00 FF B4 [3] 00 00 8B 94 [3] 00 00 E8 [4] 83 C4 0C 89 84 [3] 00 00 8D 84 [3] 00 00 B9 [4] 50 FF B4 [3]00 00 FF B4 [3]00 00 8B 94 [3]00 00 E8}
    condition:
        $ref_c2 or $ref_ecc         
}
"""
def yara_scan(raw_data):
    addresses = {}
    yara_rules = yara.compile(source=rule_source)
    matches = yara_rules.match(data=raw_data)
    for match in matches:
        for item in match.strings:
            addresses[item[1]] = hex(item[0])       #手动转为 16 进制，方便查看
    return addresses

def emotet_extract(filebuf):
    conf_dict = {}
    pe = None
    pe = pefile.PE(data=filebuf, fast_load=False)
    image_base = pe.OPTIONAL_HEADER.ImageBase       #获取载入基址，用于从 VA 转 RVA
    yara_matches = yara_scan(filebuf)

if yara_matches.get("$ref_c2"):
        delta = -5
        c2list_va_offset = int(yara_matches["$ref_c2"],16)
        c2_list_va = struct.unpack("I", filebuf[c2list_va_offset + delta : c2list_va_offset + delta + 4])[0]
        c2_list_rva = c2_list_va - image_base
        c2_list_offset = pe.get_offset_from_rva(c2_list_rva)
        key = filebuf[c2_list_offset : c2_list_offset + 4]
        presize = filebuf[c2_list_offset + 4 : c2_list_offset + 8]
        size = struct.unpack("I", presize)[0] ^ struct.unpack("I", key)[0]
        c2_list_offset += 8
        c2_list = xor_data(filebuf[c2_list_offset:], key)
        offset = 0
        while offset < size:
            ip = struct.unpack(">I", c2_list[offset : offset + 4])[0]
            c2_address = socket.inet_ntoa(struct.pack("!L", ip))        #将 32 位压缩 IPv4 地址（一个 类字节对象，长 4 个字节）转换为标准的点分十进制字符串形式（如 '123.45.67.89' ）
            port = str(struct.unpack(">H", c2_list[offset + 4 : offset + 6])[0])
            if not c2_address or not port:
                break
            conf_dict.setdefault("address", []).append(f"{c2_address}:{port}")
            c2found = True
            offset += 8

if yara_matches.get("$ref_ecc"):
        ref_ecc_offset = int(yara_matches["$ref_ecc"],16)
        delta1 = -5
        delta2 = 44
        ref_eck_rva = struct.unpack("I", filebuf[ref_ecc_offset + delta1 : ref_ecc_offset + delta1 + 4])[0] - image_base    #struct.unpack(format, buffer)，根据格式字符串 format 从缓冲区 buffer 解包，返回元祖，所以这里用[0]来提取。
        ref_ecs_rva = struct.unpack("I", filebuf[ref_ecc_offset + delta2 : ref_ecc_offset + delta2 + 4])[0] - image_base    #struct.unpack(format, buffer)，根据格式字符串 format 从缓冲区 buffer 解包，返回元祖，所以这里用[0]来提取。
        eck_offset = pe.get_offset_from_rva(ref_eck_rva)
        ecs_offset = pe.get_offset_from_rva(ref_ecs_rva)

if __name__ == "__main__":
    import sys
    with open(sys.argv[1], "rb") as f:
        file_data = f.read()
    print(emotet_extract(file_data))        #最终输出{'address': ['131.100.24.231:80', '209.59.138.75:7080', '103.8.26.103:8080', '51.38.71.0:443', '212.237.17.99:8080', '79.172.212.216:8080', '207.38.84.195:8080', '104.168.155.129:8080', '178.79.147.66:8080', '46.55.222.11:443', '103.8.26.102:8080', '192.254.71.210:443', '45.176.232.124:443', '203.114.109.124:443', '51.68.175.8:8080', '58.227.42.236:80', '45.142.114.231:8080', '217.182.143.207:443', '178.63.25.185:443', '45.118.115.99:8080', '103.75.201.2:443', '104.251.214.46:8080', '158.69.222.101:443', '81.0.236.90:443', '45.118.135.203:7080', '176.104.106.96:8080', '212.237.56.116:7080', '216.158.226.206:443', '173.212.193.249:8080', '50.116.54.215:443', '138.185.72.26:8080', '41.76.108.46:8080', '212.237.5.209:443', '107.182.225.142:8080', '195.154.133.20:443', '162.214.50.39:7080', '110.232.117.186:8080'], 'ECC ECK1': '-----BEGIN PUBLIC KEY-----\nMFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAE86M1tQ4uK/Q1Vs0KTCk+fPEQ3cuw\nTyCz+gIgzky2DB5Elr60DubJW5q9Tr2dj8/gEFs0TIIEJgLTuqzx+58sdg==\n-----END PUBLIC KEY-----', 'ECC ECS1': '-----BEGIN PUBLIC KEY-----\nMFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAEQF90tsTY3Aw9HwZ6N9y5+be9Xoov\npqHyD6F5DRTl9THosAoePIs/e5AdJiYxhmV8Gq3Zw1ysSPBghxjZdDxY+Q==\n-----END PUBLIC KEY-----'}
```

![image.png](https://shs3.b.qianxin.com/attack_forum/2022/08/attach-892d190f2553c43827b1a7987f6bfe1a35f3a5b4.png)

0x06 总结：
========

编写这种脚本时，你得知道你要什么功能，然后依照功能去找函数，找外部库。比如说我可能不知道有 pefile 这个外部库，但是我知道我需要 PE 的结构字段 Imagebase，RVA 转 FOA 这些功能，照着这些功能去搜索总能找到的。然后就是学习看官方文档，很多库在网络上的使用教程其实很少，但是官方文档描述得也不赖，而且看了这几个官方文档后发现格式排版，API 介绍，结构体对象等都有相通的地方，所以得多尝试从官方文档中找答案。

0x07 参考：
========

[CAPEv2](https://github.com/kevoreilly/CAPEv2/blob/f2ab891a278b2875c79b4f2916d086f870b54ed5/modules/processing/parsers/CAPE/Emotet.py)

[奇安信攻防社区-APT 恶意 DLL 分析及 C2 配置提取（子 DLL 篇） (butian.net)](https://forum.butian.net/share/1804)

[socket --- 底层网络接口 — Python 3.10.6 文档](https://docs.python.org/zh-cn/3/library/socket.html?highlight=socket#module-socket)

[pefile — pefile documentation](https://pefile.readthedocs.io/en/latest/modules/pefile.html)

[Welcome to YARA's documentation! — yara 4.2.1 documentation](https://yara.readthedocs.io/en/latest/)

[itertools --- 为高效循环而创建迭代器的函数 — Python 3.10.6 文档](https://docs.python.org/zh-cn/3/library/itertools.html?highlight=itertools#itertools.cycle)&lt;/i&gt;

发表于 2022-09-07 09:43:23
阅读 ( 6658 )
分类：安全开发

恶意样本自动化配置提取初探

0 条评论