In this blog post, I will describe the following three attack vectors to Microsoft Active Directory: LLMNR Poisoning, SMB Relay Attack and IPv6 Attack. These notes of mine come from the online course “Practical Ethical Hacking” by Heath Adams (aka The Cyber Mentor) where we disabled Windows Defender for demonstrative purposes.
This attack misuses the Link-Local Multicast Name Resolution (LLMNR). LLMNR is a protocol that allows name resolution without the requirement of a DNS server. It sends a request across the network asking all listening Network-Interfaces to reply if they are authoritatively known as the hostname in the query, thus providing a hostname-to-IP service.
This service can be exploited by setting up a node on the network that pretends to be whoever the query is looking for. The client requesting the information will trust whatever machine answers first, because the protocol specifies that all received responses are authoritative and trustworthy.
In our attack situation, the victim machine will request to connect to a non-existing file share and Windows will use LLMNR to identify this file server. Our attacking machine will reply to this request with Responder.py and Windows will send us the credential hash of the current user for that victim machine, thus disclosing those credentials.
Afterthought: NBT-NS is another fallback mechanism for name resolution. This service is flawed by design and should be disabled too.
SMB Relay Attack
This attack extracts a user’s credential hash from a machine and relays it to another in order to get system access on it. In order for this to work, there are two requirements. The first requirement is that SMB message signing – that is, the process of validating the source of requests against a system’s SMB services – is not enforced. The second requirement is that the user, whose credentials we relay across, is a local administrator on the machine receiving the hash.
Before starting the attack, we check which machines on the network do not enforce SMB message signing. To do this we run an nmap scan (other scan solutions are available too). If under smb2-security-mode you read “Message signing enabled but not required”, then that machine is a potential target!
The SMB Relay attack is similar to LLMNR poisoning in that we are listening for resolution requests. However, this time we won’t dump the credential hash. Instead we will reflect user authentication attempts against systems on the network. The relay mechanism we use is called ntlmrelayx.py. When ntlmrelayx.py makes a successful connection, it dumps all the local hashes on the target machine.
This attack abuses the default IPv6 configuration in Windows networks to spoof DNS replies by acting as a malicious DNS server and redirecting traffic to an attacker specified endpoint, and by then exploiting the WPAD (Windows Proxy Auto Discovery) feature with the goal of relaying credentials and authenticating to various services within the network.
The most known implementation of this attack is called mitm6. It starts with listening to the requests for an IPv6 configuration via DHCPv6 that each Windows machine makes regularly. Then mitm6 replies to these requests and sets the attacker’s IP as the default IPv6 DNS server for the victim. This is very convenient, because Windows prefers IPv6 DNS servers over their IPv4 counterpart.
Now the victim machine will start querying for the WPAD configuration of the network1. As the location of this configuration is only requested via DNS, the victim will connect to the attacker machine which will prompt authentication. The credentials will be then relayed by ntlmrelayx.py to other services in the network. ntlmrelayx.py will dump all juicy information, such as a list of domain users by group, after a successful connection.
I hope you liked this post. If you have any question, feel free to leave a comment in the comment section. Never stop learning!