Missed arguments for on_db_host_state and on_db_service_state methods.
Call these methods only when host/service state changed and pass the
old state as argument `ostate` (not sure about what `ostate` meens..)
This is a repackaging effort for the work i originally pushed in
6035. This segment of the PR provides UDP session handlers for
bind and reverse sessions, a Windows Metasm stager (really the
TCP stager with a small change), and a pair of socat payloads for
testing simple UDP shells. Netcat or any scripting language with
a sockets library is sufficient to use these sessions as they are
stateless and simple.
Testing of this PR requires rex/core #1 and rex/socket #2
The SSL testing which was being done on 6035 is backed out, left
for a later time when we can do DTLS properly.
Dnsruby provides advanced options like DNSSEC in its data format
and is a current and well supported library.
The infrastructure services - resolver, server, etc, were designed
for a standalone configuration, and carry entirely too much weight
and redundancy to implement for this context. Instead of porting
over their native resolver, update the Net::DNS subclassed Rex
Resolver to use Dnsruby data formats and method calls.
Update the Msf namespace infrastructure mixins and native server
module with new method calls and workarounds for some instance
variables having only readers without writers. Implement the Rex
ServerManager to start and stop the DNS service adding relevant
alias methods to the Rex::Proto::DNS::Server class.
Rex services are designed to be modular and lightweight, as well
as implement the sockets, threads, and other low-level interfaces.
Dnsruby's operations classes implement their own threading and
socket semantics, and do not fit with the modular mixin workflow
used throughout Framework. So while the updated resolver can be
seen as adding rubber to the tire fire, converting to dnsruby's
native classes for resolvers, servers, and caches, would be more
like adding oxy acetylene and heavy metals.
Testing:
Internal tests for resolution of different record types locally
and over pivot sessions.
Add an explicit require for the new cert_provider in framework.rb
in case it has not yet been loaded.
This should address the Travis failure on initial PR, although the
gem version in socket has not been updated, so this might take a
bit to propagate. In the end, if the dependency already gives us
this functionality by the time we call Rex::Socket::Ssl then this
commit can safely be dropped
All meterpreter Clients are created equal, and as such they all
include the PacketDispatcher mixin and call its init methods when
a passive dispatcher is needed. However, since tunneling protocols
have different requirements for implementation, the methods which
provide protocol-specific functionality need to be mixed into the
Client before it attempts to initialize the dispatcher.
Provide a dispatch_ext option in the has passed to the client on
init from the session handler which is an Array containing mixin
references which are sent to :extend calls in the :init_meterpreter
method just prior to calling :initialize_passive_dispatcher.
Each handler implementation can thus push chains of mixins to the
client in order to provide middleware specific to the tunnel. Down
the road, this should permit stacking C2 encapsulations or tunnel
protocols/permutators to create unique session transports on the
fly.
PacketDispatcher has some hardcoded assumptions about utilizing
HTTP services as the async resource. With C2 and DNS tunnels in
the pipeline, these elements need to be separated from the core
functions of async packet dispatch and moved into their own module.
This creates a new namespace for Meterpreter::HttpPacketDispatcher,
meant to be mixed in after PacketDispatcher. The module implements
only three of the original module's methods - init, shutdown, and
the :on_passive_request callback; with the first two using :super,
with the expectation of having a PacketDispatcher mixin or API
compatible namespace already in the mix.
Msf relies on Rex::Socket to create TLS certificates for services
hosted in the framework and used by some payloads. These certs are
flagged by NIDS - snort sid 1-34864 and such.
Now that Rex::Socket can accept a @@cert_provider from the Msf
namespace, a more robust generation routine can be used by all TLS
socket services, provided down from Msf to Rex, using dependencies
which Rex does not include.
This work adds the faker gem into runtime dependencies, creates an
Msf::Exploit::Remote::Ssl::CertProvider namespace, and provides
API compatible method invocations with the Rex version, but able
to generate higher entropy certs with more variables, options, etc.
This should reduce the hit rate against NIDS on the wire, reducing
pesky blue team interference until we slip up some other way. Also,
with the ability to generate different cert types, we may want to
look at extending this effort to probide a more comprehensive key
oracle to Framework and consumers.
Testing:
None yet, internal tests pending.
Travis should fail as this requires rex-socket #8.
Such as vim -i NONE. This may allow command injection via arguments.
However, you can already start an arbitrary program by setting
LocalEditor or escaping the editor.
msf > setg LocalEditor /bin/sh
LocalEditor => /bin/sh
msf > edit -i
[*] Launching /bin/sh -i
$