<aside>
During this engagement, I compromised a single host and gained SYSTEM level access. Because this was a domain-joined host, I was able to use this access to enumerate the domain. I went through all of the standard enumeration, but did not find much. There were Service Principal Names (SPNs) present within the environment, and I was able to perform a Kerberoasting attack and retrieve TGS tickets for a few accounts. I attempted to crack these with Hashcat and some of my standard wordlists and rules, but was unsuccessful at first. I ended up leaving a cracking job running overnight with a very large wordlist combined with the d3ad0ne rule that ships with Hashcat. The next morning I had a hit on one ticket and retrieved the cleartext password for a user account. This account did not give me significant access, but it did give me write access on certain file shares. I used this access to drop SCF files around the shares and left Responder going. After a while, I got a single hit, the NetNTLMv2 hash of a user. I checked through the BloodHound output and noticed that this user was actually a domain admin! Easy day from here.
</aside>
<aside>
I had tried all of my standard ways to obtain a foothold on this third engagement, and nothing had worked. I decided that I would use the Kerbrute tool to attempt to enumerate valid usernames and then, if I found any, attempt a targeted password spraying attack since I did not know the password policy and didn't want to lock any accounts out. I used the linkedin2username tool to first mashup potential usernames from the company's LinkedIn page. I combined this list with several username lists from the statistically-likely-usernames GitHub repo and, after using the userenum feature of Kerbrute, ended up with 516 valid users. I knew I had to tread carefully with password spraying, so I tried with the password Welcome2021 and got a single hit! Using this account, I ran the Python version of BloodHound from my attack host and found that all domain users had RDP access to a single box. I logged into this host and used the PowerShell tool DomainPasswordSpray to spray again. I was more confident this time around because I could a) view the password policy and b) the DomainPasswordSpray tool will remove accounts close to lockout from the target list. Being that I was authenticated within the domain, I could now spray with all domain users, which gave me significantly more targets. I tried again with the common password Fall2021 and got several hits, all for users not in my initial wordlist. I checked the rights for each of these accounts and found that one was in the Help Desk group, which had GenericAll rights over the Enterprise Key Admins group. The Enterprise Key Admins group had GenericAll privileges over a domain controller, so I added the account I controlled to this group, authenticated again, and inherited these privileges. Using these rights, I performed the Shadow Credentials attack and retrieved the NT hash for the domain controller machine account. With this NT hash, I was then able to perform a DCSync attack and retrieve the NTLM password hashes for all users in the domain because a domain controller can perform replication, which is required for DCSync.
</aside>
<aside>
</aside>
Goal:
Understand the domain: users, groups, computers, trusts, permissions.