Skip to content

Internal reconnaissance: how to tell when someone is exploring your network

How to recognize internal network reconnaissance: five signals, why they are ambiguous, and what changes when someone runs into a decoy.

9 min read
Also inenit

On a Monday morning, a server's logs show connection attempts to several systems on the company network. Some target file shares, others administrative services. Shortly after, the same account tries to log in to a server it does not normally use.

Is it an administrator checking the infrastructure? A maintenance script? Or someone who, after getting a first foothold, is working out how to move?

The answer is not necessarily obvious.

Internal reconnaissance is one of the ways a threat turns initial access into a wider knowledge of the network. Recognizing it means spotting behaviors that look normal one at a time, but mean something different when you see them together.

To see how a breach comes to be suspected in the first place, start with How to tell if your company network has been breached.

In brief

  • Internal reconnaissance is the gathering of information about the network after a first foothold: hosts, services, shares, accounts, credentials.
  • No single signal proves it. Scans, directory queries and unusual logins often have legitimate explanations, so context is needed.
  • A honeypot does not interpret behavior: it detects contact with a resource nobody should be using. It is a high-confidence signal that takes little triage.
  • It cannot see someone who never runs into a decoy, and it does not replace EDR, IDS or NDR.

What internal reconnaissance is

Internal network reconnaissance is the gathering of information about an IT environment by someone who is already in a position to explore it. A threat may try to work out:

  • which hosts and services are reachable;
  • where servers, databases and file shares sit;
  • which accounts hold interesting privileges;
  • where credentials or configuration details are stored;
  • which systems could lead to further access.

Not every attack involves extensive reconnaissance: a threat may already know its target. In many scenarios, though, exploring the environment is a useful step before deciding what to do next.

It should be told apart from lateral movement. Reconnaissance gathers information about systems and opportunities; lateral movement means actually reaching other systems. The two can alternate, but they are not synonyms.

Five signals that may point to internal reconnaissance

No single indicator proves that someone is exploring the network with hostile intent. But some behaviors deserve attention, especially when they appear together or do not match an account's usual activity.

Activity Observable signal Where to look
Host and port scanning Connections to many addresses or services in a short window IDS, NDR, network logs
Share enumeration Requests to many SMB resources or unusual paths File server logs, endpoint telemetry
User and privilege discovery Directory queries that are unusual for the account Active Directory logs, EDR
Attempts to reach other systems Authentication to hosts the account does not normally use Authentication logs, EDR
Credential and configuration hunting Unusual access to sensitive files or stores EDR, file auditing where enabled

The first three rows match techniques described in the MITRE ATT&CK Discovery tactic: Network Service Discovery (T1046), Network Share Discovery (T1135) and Account Discovery (T1087). Whether you can see them depends on the tools installed, how they are configured, and the logs actually collected.

1. Network and port scans

A threat can look for live hosts and reachable services through repeated connections to different addresses or ports. An IDS or a network detection system may recognize certain scan patterns. But a scan can also be perfectly authorized: an inventory, a vulnerability assessment, a check on a new segment. It becomes more interesting if it comes from a device that does not normally do this, or appears in an unusual sequence.

2. Share enumeration

Searching for shared folders is a way to find documents, backups and company data. Access to many shares may show up in server logs, if event logging is configured. An indexing service or an administrator can generate similar requests, so the useful question is whether that account has a plausible reason to look for them.

3. Looking up users, groups and privileges

In a Windows environment, a threat can query Active Directory to learn which accounts exist and which groups they belong to. Many queries are ordinary; others look unusual for the device or identity running them. Spotting them takes adequate visibility and, often, correlation with other activity.

4. Authentication to unusual systems

An account that starts trying to reach many servers can point to exploration, or to the start of lateral movement. But an administrator may work on several systems during an emergency, and a new application can change the usual access paths. That is why the sequence matters: what happened before, which resources were reached, which privileges were used.

5. Hunting for credentials and configuration files

Saved credentials, configuration files and technical documents can help a threat find other systems. Endpoint detection tools can recognize some credential-access techniques, and file auditing can add information where available. Not every file read is suspicious and not every read is logged: a technical document opened by an administrator is normal, while the same document searched for by a compromised account after unusual logins means something different.

Why it is hard to tell reconnaissance from normal IT work

Many operations threats use are also everyday administration tools. Scanning a network, querying a directory or logging in to a server is not malicious in itself. To give these events a meaning you need extra information: who acted, from which system, at what time, in what sequence.

A well-configured EDR, IDS or NDR can recognize activity that would otherwise go unnoticed. The problem, especially in organizations with limited IT resources, is deciding which signals need immediate action. Detecting a suspicious event and giving it a meaning are two different problems, and that is where deception brings a complementary approach.

What changes when someone runs into a decoy

Imagine that, while looking for interesting systems, someone finds a server that seems to be dedicated to backups. It is believable, answers connections and offers working services, but it holds no real backups and does no operational job. It is a honeypot.

If someone tries to access it, the event means something specific: no user or business process should be using that resource. An authorized scan or an operational mistake can still reach a decoy, so not every connection proves an intrusion. But the legitimate explanations are far fewer than on a production server.

A honeypot does not have to decide whether an activity is unusual compared with normal traffic: it detects interaction with a target that should not be used. The result is a high-confidence signal, particularly useful when there is little time to check alerts. The limit is just as clear: if the threat never runs into the decoy, the honeypot does not detect it.

How Hoxey increases the chances of detection

Hoxey uses high-interaction honeypots and widens the deception surface with elements placed on endpoints, to create more occasions where hostile activity can produce a meaningful signal.

  • High-interaction honeypots. Decoys are complete systems, with plausible services and fictional data: whoever reaches one finds a believable environment, not a mock login screen.
  • Baits. Lures such as credentials or configuration fragments that point to an apparently interesting resource and can lead a threat to a decoy. Baits are inert files: there is no agent to install.
  • Tripwires. Artifacts set up to produce an observable request when they are triggered, for example when they are opened or indexed. They make an interaction visible even when the threat uses no credentials and connects to no server.
  • Deception network. Once a threat reaches a decoy, it can run into further decoy resources, all hosted in Hoxey's environment and not in the customer's network. The alert that matters has already fired at first contact.

The risk of putting an attackable system on your network

The historical limit of interactive honeypots is the risk that a compromised decoy becomes a starting point for reaching the rest of the network: what is known as a pivot.

Hoxey hosts its decoys in its own cloud, outside the customer's LAN. A link component sets up an encrypted connection and makes the decoy reachable as if it were a local resource. The decoy can answer the requests it receives, but it cannot start new connections toward the company network through that link. The principle is to stop the decoy environment from becoming a path into real systems. The mechanism is described in the Hoxey Security Model, and the topic is explored further in Can a honeypot be used against me?.

What a honeypot cannot detect

Deception does not give complete visibility over everything happening inside. A threat can explore real systems, reach data or move between hosts without ever interacting with a decoy, a bait or a tripwire. In those cases detection depends on other controls: EDR, IDS, NDR, authentication logs and monitoring suited to the environment. A honeypot does not replace them, it adds a different kind of signal.

To go further on how the two approaches differ, see also Honeypot vs IDS: what each one actually detects, and why you may need both.

How to assess your ability to recognize internal reconnaissance

Before adding new tools, an IT manager can ask five questions:

  1. Do we see internal scans? Can we tell when a host queries many addresses or services?
  2. Do we have visibility of unusual authentication? Can we spot an account trying to reach systems it does not normally use?
  3. Are we collecting the events we need? Do server, directory and endpoint logs really allow us to rebuild a sequence of activity?
  4. Who checks the signals? Is there a person or a service in charge of assessing the events that need attention?
  5. Do we have believable decoys? Could a threat exploring the environment run into something with no legitimate use, which produces a high-confidence signal?

The first four questions are about visibility and the capacity to analyze. The fifth introduces a different way of recognizing hostile activity.

Deception does not make everything on the network visible. It makes what happens when someone meets a decoy far more meaningful.

Frequently asked questions

What is internal reconnaissance on a company network?

It is the gathering of information about hosts, services, shares, accounts and credentials by someone who already has a first foothold and wants to work out how to move.

What is the difference between internal reconnaissance and lateral movement?

Reconnaissance gathers information about systems and available opportunities. Lateral movement means actually reaching other systems. The two can alternate, but they are not the same thing.

How can you tell that internal reconnaissance is happening?

There is no single indicator. You look for scans across many hosts, requests to many shares, unusual directory queries, logins to systems an account does not normally use, and unusual access to sensitive files. They need to be read together and in context, because each one can have a legitimate explanation.

Can a honeypot detect internal reconnaissance?

Yes, when whoever is exploring the network runs into a decoy: a decoy server, a bait or a tripwire. The signal is high-confidence because no user or business process should interact with that resource. It does not detect someone who never runs into one.

Does a honeypot replace EDR, IDS or NDR?

No. It adds a different kind of signal. Activity that never touches a decoy remains the job of the other controls.

That is the capability Hoxey adds to intrusion detection.

Send threat actorsfrom your network into ours.

From €145 per decoy per month. Free trial available, no hardware, no setup fees.

Join The Hive

Threat intel from our own decoys, first-hand. Not another newsletter.

  • Intel from our network. What our decoys catch: the credentials threats try, the tools they bring, the techniques that are new. First-hand, not recycled.
  • Alerts, when it matters. When something we see is worth acting on, you hear it from us, with what to do.
  • Attack stories. The story behind the intel: a real attack, narrated step by step from the offensive side.

Low volume, high quality.

Plus occasional Hoxey news, and we remember your visits. Unsubscribe anytime.

Get in touch

Write to us here and we'll email you back.

We'll reply by email. Privacy policy Prefer a call? Book one