Dan Farmer noted that the RAKP reference in the title was not correct
and that RAKP is a separate issue and protocol implementation than
the use of Cipher Zero to perform an authentication bypass.
Cosmetic only change
console1:
```
msf> use auxiliary/scanner/http/open_proxy
msf auxiliary(open_proxy) > set RHOSTS 192.168.0.4
msf auxiliary(open_proxy) > set RPORT 8888
msf auxiliary(open_proxy) > run
< the connection never times out >
```
console2:
```
$ nc -vlp 8888
Listening on [0.0.0.0] (family 0, port 8888)
Connection from [192.168.0.4] port 8888 [tcp/*] accepted (family 2, sport 43245)
GET http://209.85.148.147/ HTTP/1.1
Host: 209.85.148.147
Connection: close
User-Agent: user_agent
Accept-Encoding: *
Accept-Charset: ISO-8859-1,UTF-8;q=0.7,*;q=0.7
Cache-Control: no
Accept-Language: de,en;q=0.7,en-us;q=0.3
```
After the patch, requests timeout after 10 seconds:
```
msf auxiliary(open_proxy) > run
[*] Scanned 1 of 1 hosts (100% complete)
[*] Auxiliary module execution completed
```
Honour the user supplied settings
Abort a host on connection error
Check a 200 response for some appropriate data
Let datastore validation handle things like options being present
Be more verbose if needed
Use the HTTPClient more appropriately
Anemone::Page is not a Rex HTTP request/response, and uses the
:cookies method to return an array of cookies.
This resolves the method naming error, though it does break with
Rex naming convention since Anemone still uses a lot non-Rex
methods for working with pages/traffic.
Instead of editing the datastore['PATH'], use normalize_uri.
Since the purpose of this module is quite fuzz-like, I didn't want to
apply the normalize_uri to the whole uri -- the original code merely
applied to datastore['PATH'] (which seems like it should be
datastore['URI'] really) and then added on a bunch of other stuff to
test for traversals.
If the user wants to use downcased or mixed case HTTP methods, heck,
more power to them. If it doesn't work, it doesn't work. No other HTTP
module makes this call.
Looks like copypasta artifacts. DBUSER and DBPASS aren't ever set as
options in the module, and the module doesn't include MC's
Exploit::ORACLE mixin. It's also from four years ago and doesn't
report_auth or anything useful like that, but that's out of scope for
this branch.
This module was written in an era where the defaults for bruteforcing
included a lot of lock-inducing behavior, thus, it was quite serious
about setting datastore options directly. Also, there was apparently a
bug in USER_AS_PASS that this module attempted to avoid by setting the
datastore directly, rather than fixing the bug directly. As far as I
know, this bug has been long since resolved.
The attached updates changes smb_login to detect if the newly discovered user is an administrator. It is based on code from Brandon McCann "zeknox" submitted in PR #1373, the associated changes, and the newer PR #2656.
The changes should correct a few issues with PR #1373 and #2656 and address Redmine bug #8773.
Specifically it:
- Fixes the admin detection code by using simple.disconnect(<share>) instead of disconnect()
- Adds support for detecting if the remote host will allow connects using any domain name when one of the new status codes is returned
- Dealt with the issue in PR #2656 where the username was prefixed with a '\'
Verification
Be connected to a database
Run this against a machine with a known user and admin user
See that the admin user is reported correctly
See that the non-admin user is reported correctly
Check the output of creds
Select a target that requires a domain in order to authenticate
In the stored credentials, with CHECK_ADMIN enabled, see that the domain name is, in fact, preserved in the reporting
To validate that the remote domain ignores domain value use the following command from a windows system:
net use \\<hostip>\admin$ /user:<random_value>\<username> <password>
Christian Mehlmauer
Rob Fuller
jvazquez-r7
Meatballs
attackdebris
HD Moore
scriptjunkie
Tod Beardsley
William Vu
RageLtMan
joev
sappirate
Arnaud SOULLIE
Tom Sellers
sinn3r