OT / ICS security assessment (passive-first)
When it applies
The engagement touches an industrial environment — PLC/RTU controllers, SCADA/DCS, HMIs,
engineering workstations, historians, jump hosts, or the IT/OT boundary (firewall rules, data
diodes). A misstep here can cause physical harm, not just downtime. Load tradecraft-scope-roe
and require the written authorization to spell out: the site and network segments, whether active
scanning is allowed, and whether writing coils/registers is allowed — the default for both is
NO. Work passive-first; nothing touches a PLC's state until the scope explicitly says so.
Why it works
Industrial protocols (Modbus, S7comm, DNP3, EtherNet/IP) predate hostile networks: they run in
cleartext with no authentication (CWE-306), so anyone who can reach the port can read process
state — and usually write it. That's exactly why the assessment itself is the risk: a high-rate
scan or a stray write can trip a control loop or a safety instrumented system (SIS). Mirrored
traffic and offline artifacts reveal the same exposure with zero packets injected.
Iron rules (MUST)
- Never write coils/registers to a PLC unless the authorization explicitly permits it.
- No high-rate scanning of production OT segments — low rate, maintenance window only.
- Never touch SIS paths or anything in a safety loop.
- Prefer, in order: read-only identification → traffic mirroring → offline firmware/config analysis.
- An anomaly or unexpected device response → stop immediately and notify the site contact.
Method
- Zone and inventory (paper-first). Sketch the Purdue levels L0–L5: field devices → basic
control → supervisory → site DMZ → enterprise. Build the asset list — PLC/RTU/HMI/engineering
workstation/historian/jump host — and the expected protocol/port baseline for the authorized
segments only (identification, not exploitation): Modbus/TCP 502, S7comm 102,
EtherNet/IP 44818, DNP3 20000.
- Passive capture. Take a SPAN/mirror-port PCAP from an authorized tap:
tshark -i <iface> -w ot-mirror.pcap. Dissect it with Wireshark's ICS dissectors (modbus,
s7comm, enip, dnp3) — you learn the live devices, the masters/slaves, the function codes in use,
and any cleartext credentials without sending a single packet. A proprietary protocol on the
wire goes to reverse-eng-protocol.
- Offline audit. Review exported controller configuration and engineering projects (TIA
Portal, RSLogix/Studio 5000 exports) offline. Record default credentials and unauthenticated
cleartext protocols (Modbus has none) as findings — observe and document; change nothing.
- Limited active checks — only if the scope says yes. During the agreed maintenance window,
at low rate (
nmap -Pn -n --max-rate 20 -p 502,102,44818,20000 <host>, plus targeted NSE
identification such as --script modicon-info,s7-info,enip-info), prefer read-only function
codes, capture evidence for every step, and stop + escalate at the first anomaly.
- Firmware and patch surface. Map controller firmware versions to known CVEs (never blind-flash
firmware). Pull images for offline analysis with
reverse-eng-firmware and reverse-eng-binary-triage
rather than probing the live controller.
- Report with physical consequence. Every OT finding states its process/physical impact
(what an attacker could make the plant do), and distinguishes remotely exploitable issues from
ones requiring physical or control-network access. Lateral movement into the IT side follows
network-pivoting-tunneling and the ad domain skills as usual.
Gotchas
- A port scan is an active act on OT. Some PLCs fault or reboot under ordinary scan rates —
mirrored traffic first, always. Don't run default-parameter web/network scanners against OT
segments; the profiles are wrong and the risk is real.
- "No response" may mean you hurt something, not that nothing's there — if a device goes quiet
mid-test, stop and notify; check with the site before resuming.
- Most ICS tooling assumes an isolated lab network — stage and rehearse tools in a lab VLAN or
on a mirrored feed, not on the production segment.
- Default creds and no-auth protocols are findings, not invitations — log them, prove reachability
minimally, never exercise a write "to demonstrate".
- Confirm the emergency contact and rollback plan exist before any active phase — including the
maintenance window and who can halt the test.
Verify success
You can produce: the Purdue-zone sketch with a reconciled asset inventory, the observed
protocol/port baseline from passive capture, a findings list with physical-impact framing
(remote vs physical reachability stated per finding), and — for any authorized active step — a
step-by-step evidence trail showing read-only behavior and zero unplanned state changes.
References
NIST SP 800-82 (ICS security); ATT&CK for ICS; Wireshark ICS dissector docs; reverse-eng-firmware.
Portions adapted from reverse-skill by zhaoxuya520, MIT License.
1---2name: hardware-ot-ics3description: >4 Assess OT/ICS environments — PLC, SCADA, DCS, HMI — safely and passive-first: Purdue-model5 zoning, industrial protocol discovery (Modbus, S7, DNP3, EtherNet/IP), mirrored-traffic analysis,6 and read-only verification. Load for an authorized engagement touching industrial control7 networks, engineering workstations, historians, or IT/OT boundaries. Signals: ports 502, 102,8 44818, 20000; Modbus/S7comm banners; PLC/RTU inventory requests.9domain: hardware10type: methodology11stability: learning12modes: [pentest, defense]13severity: high14mitre: [T0846, T0842]15cwe: [CWE-306]16tools: [wireshark, tshark, nmap, binwalk, ghidra]17schema_version: 118metadata:19 source: https://skillmd.com/skills/noorqureshi/hardware-ot-ics20---2122# OT / ICS security assessment (passive-first)2324## When it applies25The engagement touches an industrial environment — PLC/RTU controllers, SCADA/DCS, HMIs,26engineering workstations, historians, jump hosts, or the IT/OT boundary (firewall rules, data27diodes). **A misstep here can cause physical harm, not just downtime.** Load `tradecraft-scope-roe`28and require the written authorization to spell out: the site and network segments, whether active29scanning is allowed, and whether writing coils/registers is allowed — **the default for both is30NO**. Work passive-first; nothing touches a PLC's state until the scope explicitly says so.3132## Why it works33Industrial protocols (Modbus, S7comm, DNP3, EtherNet/IP) predate hostile networks: they run in34cleartext with no authentication (`CWE-306`), so anyone who can reach the port can read process35state — and usually write it. That's exactly why the assessment itself is the risk: a high-rate36scan or a stray write can trip a control loop or a safety instrumented system (SIS). Mirrored37traffic and offline artifacts reveal the same exposure with zero packets injected.3839## Iron rules (MUST)40- **Never write coils/registers** to a PLC unless the authorization explicitly permits it.41- **No high-rate scanning** of production OT segments — low rate, maintenance window only.42- **Never touch SIS paths** or anything in a safety loop.43- Prefer, in order: read-only identification → traffic mirroring → offline firmware/config analysis.44- An anomaly or unexpected device response → **stop immediately and notify** the site contact.4546## Method471. **Zone and inventory (paper-first).** Sketch the Purdue levels L0–L5: field devices → basic48 control → supervisory → site DMZ → enterprise. Build the asset list — PLC/RTU/HMI/engineering49 workstation/historian/jump host — and the expected protocol/port baseline for the authorized50 segments only (identification, not exploitation): Modbus/TCP **502**, S7comm **102**,51 EtherNet/IP **44818**, DNP3 **20000**.522. **Passive capture.** Take a SPAN/mirror-port PCAP from an authorized tap:53 `tshark -i <iface> -w ot-mirror.pcap`. Dissect it with Wireshark's ICS dissectors (modbus,54 s7comm, enip, dnp3) — you learn the live devices, the masters/slaves, the function codes in use,55 and any cleartext credentials without sending a single packet. A proprietary protocol on the56 wire goes to `reverse-eng-protocol`.573. **Offline audit.** Review exported controller configuration and engineering projects (TIA58 Portal, RSLogix/Studio 5000 exports) offline. Record default credentials and unauthenticated59 cleartext protocols (Modbus has none) as findings — **observe and document; change nothing**.604. **Limited active checks — only if the scope says yes.** During the agreed maintenance window,61 at low rate (`nmap -Pn -n --max-rate 20 -p 502,102,44818,20000 <host>`, plus targeted NSE62 identification such as `--script modicon-info,s7-info,enip-info`), prefer read-only function63 codes, capture evidence for every step, and stop + escalate at the first anomaly.645. **Firmware and patch surface.** Map controller firmware versions to known CVEs (never blind-flash65 firmware). Pull images for offline analysis with `reverse-eng-firmware` and `reverse-eng-binary-triage`66 rather than probing the live controller.676. **Report with physical consequence.** Every OT finding states its process/physical impact68 (what an attacker could make the plant *do*), and distinguishes remotely exploitable issues from69 ones requiring physical or control-network access. Lateral movement into the IT side follows70 `network-pivoting-tunneling` and the `ad` domain skills as usual.7172## Gotchas73- **A port scan is an active act on OT.** Some PLCs fault or reboot under ordinary scan rates —74 mirrored traffic first, always. Don't run default-parameter web/network scanners against OT75 segments; the profiles are wrong and the risk is real.76- **"No response" may mean you hurt something, not that nothing's there** — if a device goes quiet77 mid-test, stop and notify; check with the site before resuming.78- **Most ICS tooling assumes an isolated lab network** — stage and rehearse tools in a lab VLAN or79 on a mirrored feed, not on the production segment.80- **Default creds and no-auth protocols are findings, not invitations** — log them, prove reachability81 minimally, never exercise a write "to demonstrate".82- **Confirm the emergency contact and rollback plan exist before any active phase** — including the83 maintenance window and who can halt the test.8485## Verify success86You can produce: the Purdue-zone sketch with a reconciled asset inventory, the observed87protocol/port baseline from passive capture, a findings list with physical-impact framing88(remote vs physical reachability stated per finding), and — for any authorized active step — a89step-by-step evidence trail showing read-only behavior and zero unplanned state changes.9091## References92NIST SP 800-82 (ICS security); ATT&CK for ICS; Wireshark ICS dissector docs; `reverse-eng-firmware`.9394---95_Portions adapted from [reverse-skill](https://github.com/zhaoxuya520/reverse-skill) by zhaoxuya520, MIT License._