Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is LLMNR and NBT-NS Poisoning?

LLMNR and NBT-NS poisoning answers a Windows host’s fallback name lookups with the attacker’s own address, collecting NTLMv2 authentication from users who simply mistyped a share name.

Illustration for the cybersecurity term LLMNR and NBT-NS Poisoning
Home / Blog / Cybersecurity Terms / What Is LLMNR and NBT-NS Poisoning?
17 November 2026

What Is LLMNR and NBT-NS Poisoning?

What Is LLMNR and NBT-NS Poisoning?

Ask any internal penetration tester what they run in the first five minutes of an engagement and the answer is usually Responder. LLMNR and NBT-NS poisoning is the reason: it costs nothing, it needs no vulnerability, and on a typical Windows network it produces authentication material from real user accounts within minutes of plugging in.

The Protocols Nobody Meant to Leave On

When a Windows machine cannot resolve a name through DNS, it does not give up. It falls back to Link-Local Multicast Name Resolution, defined in RFC 4795, which broadcasts the query to UDP port 5355 on the local segment, and then to NetBIOS Name Service on UDP port 137. Apple devices add mDNS on UDP 5353. All three ask the same question of everyone present: “does anyone here know who fileserv01 is?”

None of them authenticate the answer. Whoever replies first is believed.

The queries happen constantly for entirely innocent reasons. A user mistypes a share path. A login script references a server decommissioned two years ago. A printer driver looks up a WINS name. An Outlook profile still points at a retired mailbox host. Every one of those failures becomes an opportunity.

From a Reply to a Credential

The attack unfolds in a few seconds:

  1. The victim’s DNS lookup for a non-existent name fails.
  2. The victim multicasts an LLMNR query.
  3. The attacker, running Responder, answers “that is me” with their own IP address.
  4. The victim connects—usually over SMB—and the attacker’s rogue server demands authentication.
  5. Windows automatically sends the logged-on user’s NTLMv2 challenge-response without prompting, because the host believes it is talking to a domain resource.

The attacker now holds a NetNTLMv2 response, which is not the password hash itself but can be cracked offline with Hashcat in mode 5600. A dictionary attack against a company’s typical password pattern recovers a usable share of them, particularly for service accounts whose passwords were set once in 2017.

Cracking is the slow option. The faster path is NTLM relay: pass the authentication straight on to another host that does not require SMB signing, or to LDAP on a domain controller, and act as the victim without ever learning their password. Impacket’s ntlmrelayx.py paired with Responder in analyse mode is the standard combination.

Why It Survives in 2026

These protocols remain enabled by default on Windows clients, and disabling them is a group policy change nobody is assigned to make. Network teams often assume the issue was handled during a Windows 10 rollout; it rarely was. Flat internal networks make it worse, because a single compromised workstation, or a visitor port in a meeting room, reaches every host that might make a failed lookup.

There is also a sense that “only hashes” leak. In an Active Directory environment, a service account’s NetNTLMv2 response is frequently the first link in a chain that ends in Domain Admin via Kerberoasting or delegation abuse.

Shutting It Down

Test the change rather than assuming it worked. Run Responder in analysis mode, which listens without answering, and confirm the queries have stopped.

Detection

A simple and effective tripwire is to look for any LLMNR or NBT-NS response on a network where the protocols should be disabled—there should be none. A second approach is to seed a hostname that does not exist anywhere, then alert when any host attempts to authenticate to something claiming that name. Network detection and response platforms and honeypot listeners both handle this well, and either gives you a high-confidence alert with almost no false positives.

At Privia Security, we start internal Penetration Test Services with exactly this technique, because the result tells you a great deal about an estate’s credential hygiene in the first few minutes.

FAQ – Frequently Asked Questions

Will Disabling LLMNR Break Anything?

Rarely, provided DNS is healthy. LLMNR only activates when a DNS lookup fails, so correctly registered hosts are unaffected. Problems appear in environments that quietly depend on name resolution for machines absent from DNS, such as isolated lab segments or ad hoc peer-to-peer shares. Audit queries with Responder in analysis mode first, add the missing DNS records, then disable the protocol.

Is a Cracked NetNTLMv2 Response the Same as a Password Hash?

No. The captured value is a challenge-response calculated from the NT hash, not the hash itself, so it cannot be replayed with pass-the-hash. It can be cracked offline to recover the plaintext password, or relayed live to another service during the short window in which it is valid. Both outcomes are serious, which is why capture should be treated as compromise of that account.

Does Disabling NTLM Entirely Solve the Problem?

It closes the most damaging outcome but not the technique. With NTLM disabled, a poisoned response can still redirect traffic and may capture other credential types or enable interception. Most estates also find full NTLM removal a multi-year project because of legacy applications and appliances. Disabling the name resolution fallbacks is faster, lower risk, and addresses the root cause directly.

How Often Should We Re-Test for This?

After every significant change to group policy, imaging, or network design, and at least annually. Group policy exceptions accumulate, new subnets are built from older templates, and acquired companies arrive with their own defaults. A ten-minute Responder run in analysis mode on each VLAN gives a definitive answer and should be part of routine internal assurance rather than only an annual test.

Related Terms

You May Be Interested In These

What Is Envelope Encryption?

Cybersecurity Terms

17 November 2026

What Is Envelope Encryption?

Read More
What Is VERBIS?

Cybersecurity Terms

17 November 2026

What Is VERBIS?

Read More
What Are Zones and Conduits?

Cybersecurity Terms

17 November 2026

What Are Zones and Conduits?

Read More