Reporting Tools Vulnerability Analysis

Frida-Wshook – Script Analysis Tool Based On

Frida-Wshook - Script Analysis Tool Based On

frida-wshook is an analysis and instrumentation tool which uses to hook common functions often used by malicious script files which are run using WScript/CScript.

The tool intercepts Windows API functions and doesn’t implement function stubs or proxies within the targeted scripting language. This allows it to support analyzing a few different script types such as:
  • .js (JScript)
  • .vbs (VBScript)
  • .wsf (WSFile) (Initial support/testing. – Does not support specific jobs)
By default script files are run using cscript.exe and will output:
  • COM ProjIds
  • DNS Requests
  • Shell Commands
  • Network Requests
Warning!!! Ensure that you run any malicious scripts on a dedicated analysis system. Ideally, a VM with snapshots so you can revert if a script gets away from you and you need to reset the system.
Although common methods have been hooked, Windows provides numerous APIs which allow developers to interact with a network, file system and execute commands. So it is entirely possible to encounter scripts leveraging uncommon APIs for these functions.

Install & Setup

pip install frida
  • Clone (or download) the frida-wshook repository.

Supported OS
frida-wshook has been tested on Windows 10 and Windows 7 and should work on any Windows 7 + environment. On x64 systems CScript is loaded from the C:WindowsSysWow64 directory.
It may work on WindowsXP, but I suspect that CScript may use the legacy API calls and would bypass the instrumentation.

The script supports a number of optional commandline arguments that allow you to control what APIs the scripting host can call.

usage: [-h] [--debug] [--disable_dns] [--disable_com_init]
                       [--enable_shell] [--disable_net]
                       script your friendly WSH Hooker
positional arguments:
script Path to target .js/.vbs file

optional arguments:
-h, --help show this help message and exit
--debug Output debug info
--disable_dns Disable DNS Requests
--disable_com_init Disable COM Object Id Lookup
--enable_shell Enable Shell Commands
--disable_net Disable Network Requests

Analyze a script with the default parameters:

python bad.js

Enable verbose debugging:

python --debug bad.js

Enable shell (execute) commands:

python --enable_shell bad.vbs

Disable WSASend:

python --disable_net bad.vbs

Check what ProgIds the script uses:

python --disable_com_init bad.vbs

Hooked Functions

Known Issues

  • Network responses are not captured
  • Disabling Object Lookup can cause the script to only output the first ProgId…Malware QA can be lacking.
  • WSF files with a specific job to target currently isn’t supported


  • Change GetAddrInfoExW to use .replace instead of .attach
  • Add additional tracing and hooks to cover more APIs
  • Look at bypassing common anti-analysis techniques found in scripts (sleeps etc)
  • Update and improve network request hooking (ie: currently it captures requests, but not responses)

Feedback / Help
Any questions, comments or requests you can find us on twitter: @seanmw or @herrcore