HackBlue

Tools · Network tools · 24 / 24

NetworkMiner: when it is the wrong tool

NetworkMiner is the right first look for hosts, cleartext files, and cleartext credentials in a pcap you already have. It is the wrong tool for a single frame, a week of sensors, or a document that never left TLS.

14 min read · practice only with authorization

Questions that belong elsewhere

A TCP flag, a sequence number, or one option in frame 441 is tcpdump or tshark. A count of S0 connections over a day of sensor logs is Zeek or the SIEM that stores it. NetworkMiner reads packets, not a folder of conn.log. A continuous watch for the inventory-job fan-out belongs on the sensor, not in a GUI you open after the incident.

A file transferred inside TLS will not come back as that file. Certificate names and SNI may be visible in a current build. The body is not. Decrypting it requires a key log the owner already chose to provide, and even then the field work is tshark's, not a second guess in the Files tab. High-rate live capture is also the wrong job for a case GUI. Capture with tcpdump under the scope, then open the bounded file.

  • Frame-level questions go to tshark or tcpdump.
  • Multi-day detection goes to Zeek on the sensor.
  • Log-only cases have nothing for NetworkMiner to open.
  • TLS bodies stay out of the Files tab. Say so and move on.

A worked defensive case

Three questions arrive on the files incident. Which lab hosts are in the pcap, and did any cleartext file move? NetworkMiner answers from Hosts and Files. Did the 10:14 session finish the TCP handshake before the client reset? tcpdump on that pair answers, and NetworkMiner was never going to show Flags [R.]. Should the sensor page the next time 10.8.0.20 fans out S0? That is a Zeek notice on the lab sensor, tested against a later log, not a saved GUI session.

You write the split into the ticket so the next person does not re-open NetworkMiner for the reset or export conn.log into a pcap. Each tool keeps the question it can actually see.

Where this is documented

Netresec positions NetworkMiner as a network forensics tool for captured traffic and extracted artifacts. tcpdump(1), tshark(1), and Zeek's logs cover the questions this GUI does not. The scope still decides which file any of them may open.

Use a host-only lab network, or a capture the owner already handed you. These commands read that file or record a single lab host. They are not a scan of the public internet.

Commands

The frame-level question, answered from the same pcap
tcpdump -nn -tttt -r /lab/evidence.pcap -c 30 'host 10.8.0.20 and host 10.8.0.40 and tcp port 443'
The log question, answered without converting logs into a pcap
zeek-cut -m ts uid id.orig_h id.resp_p conn_state history < /lab/zeek/conn.log

Diagram

Pick the tool that can see the thing you were asked

Same incident, three questions

  • Hosts and cleartext files

    NetworkMiner. The pcap is the input. You leave with an inventory, hashes, and redacted credential notes.

  • Flags and one frame

    tcpdump or tshark. A session row will not show Flags [S.] or a frame number.

  • Tomorrow's detection

    Zeek on the sensor. A GUI session does not watch the lab after you close it.

Check yourself

1. The question is which cleartext files moved between lab hosts in this pcap. Which tool do you open?

2. The question is whether the client sent RST after the handshake. Why is NetworkMiner the wrong place?

3. You have a week of conn.log and no pcap. What can NetworkMiner add?