Guide · Agent skill
How to use the OT / ICS Security Assessment agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The ot-ics skill is a set of written instructions that tells an AI agent how to carry out an authorized security review of industrial control systems such as PLCs, SCADA servers and HMIs. It is aimed at security professionals and engineers who work on factory, utility or similar OT networks. After this guide you will know what the skill changes about the agent's behaviour, how to install it, how to ask for it in plain language, and which safety limits it enforces.
OT / ICS Security Assessmentzhaoxuya520/reverse-skill · MIT Download the skill Quick start
- 1Get the skill files. Download the ZIP from this page and unzip it so you have the ot-ics folder containing SKILL.md and references/ot-safe-assessment.md.
- 2Place the skill folder. Copy ot-ics into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder to limit it to that project. Alternatively, upload it under Settings → Capabilities → Skills in the Claude apps.
- 3Install the companion tools. Install the analysis tools you plan to use, such as Wireshark with OT dissectors, Nmap NSE, binwalk and Ghidra. Note that the companion skills firmware-pentest and protocol-reverse are not included and must be obtained separately.
- 4Write down your authorization. Record the sites, network segments, emergency contacts and whether active scanning or register writes are allowed. The skill expects this before any active step.
- 5Ask for a passive-first review. Start a request that names the authorized scope and asks for passive, offline analysis first.
- 6Review findings with process impact. Check that each finding describes its physical or process impact and separates remote exploitation from physical access.
What does the ot-ics skill make the agent do?
An agent skill is a folder containing instructions (here, a SKILL.md file plus a reference checklist) that an AI agent reads when a task matches the skill's description. This skill covers authorized OT/ICS assessment: Purdue model zoning, PLC and SCADA exposure, industrial protocol discovery and safe, passive-first evaluation.
The skill pushes the agent to work in a fixed order: sketch the Purdue zones, build an asset inventory, analyse captured traffic and exported configuration files offline, and only then consider light, read-only checks if written authorization allows them.
- Map zones from field devices (L0) up to enterprise (L5), including site DMZ.
- Inventory assets such as PLCs, RTUs, HMIs, engineering workstations, historians and jump hosts.
- Review mirrored (SPAN) packet captures and offline configuration or project exports.
- Record default credentials and unencrypted protocols such as Modbus (no authentication) as findings rather than changing values.
- Describe physical or process impact in each finding.
Install the skill from the ZIP or the Claude apps
Download the ZIP from this page and unzip it. For a personal install, place the ot-ics folder in ~/.claude/skills/. To scope it to one project, place it in that project's .claude/skills/ folder. You can also upload the skill in the Claude apps under Settings → Capabilities → Skills.
The package contains references/ot-safe-assessment.md and SKILL.md. The skill runs no scripts and makes no network calls itself, but it refers to companion skills and tools that you must install separately (see the requirements section).
- Personal: ~/.claude/skills/ot-ics/
- Project: .claude/skills/ot-ics/ inside your project
- Claude apps: Settings → Capabilities → Skills, then upload
~/.claude/skills/ot-ics/Trigger the skill with natural requests
You do not need a special command. Describe an authorized OT task in plain language that mentions industrial equipment or the assessment type, and the agent can match it to the skill's description. Including your authorization scope in the request helps the agent follow the skill's rules from the start.
The skill's own workflow asks for written authorization to name the sites, network segments, and whether active scanning or register writes are permitted. Put those details in your first message.
- "Using the ot-ics skill, plan a passive-first Purdue zoning review of site A, scope limited to 10.20.0.0/24 under our signed authorization."
- "Help me inventory PLCs and HMIs from this SPAN capture. Analyse offline only."
- "Review this exported TIA Portal project for default passwords and list findings with process impact."
What the skill needs before you start
The skill itself is only instructions. Its workflow refers to several things that are not included in the files: the companion skills firmware-pentest and protocol-reverse, and a field-journal precedent file. You also need to install industrial analysis tools yourself if you use them.
The source names Wireshark with OT dissectors, Nmap NSE, binwalk and Ghidra as the kinds of tools the workflow uses. Vendor engineering software and commercial platforms such as Claroty or Nozomi are mentioned as optional and are not bundled.
- Companion skills: firmware-pentest, protocol-reverse (not included in these files).
- Tools you install separately: Wireshark with OT dissectors, Nmap NSE, binwalk, Ghidra.
- Optional vendor or commercial tools: PLC engineering software, Claroty, Nozomi.
- Written authorization covering sites, segments and permitted activity.
Safety rules the skill enforces
The skill is built around caution because mistakes on industrial networks can cause physical harm and outages. It states that the agent must not write coils or registers on a PLC, must not run high-rate scans of production OT, and must not interrupt safety instrumented system (SIS) paths unless that is explicitly allowed.
The skill's reference checklist asks you to confirm scope and emergency contacts, keep a maintenance window and rollback plan, prefer traffic mirroring over port scanning, stop and report serious findings, and separate remotely exploitable issues from those needing physical access.
- Default to passive and read-only work; no PLC writes before the ready-for-act gate.
- Prefer mirrored traffic over active port scans.
- Use low-rate identification only inside a maintenance window.
- Stop and notify when a high-risk finding appears.
- Do not use ordinary web-scanner defaults against OT systems.
Example workflow from start to finish
A typical run starts with scope and zoning, moves to offline analysis, and only then to limited active checks. The skill lists read-only function codes as the first choice in active phases and asks for evidence at every step.
Each finding should note the affected asset, the observed condition, and its physical or process impact. Keep a checklist or journal so the assessment can be reviewed later.
- Phase 1: Purdue sketch, asset list, protocol and port baseline for authorized segments only.
- Phase 2: Analyse SPAN/mirror PCAP and offline configuration exports.
- Phase 3: Limited active work only if authorized, low rate, read-only where possible.
- Phase 4: Map controller firmware versions to known CVEs without flashing firmware.
What people use it for
- Building a Purdue model zone diagram for a site from network documentation and captures.
- Inventorying PLCs, RTUs, HMIs and historians from a mirrored traffic capture.
- Auditing an exported engineering project for default passwords or unencrypted protocols, offline.
- Creating a safety-first checklist before any active testing is scheduled in a maintenance window.
- Reviewing an IT/OT boundary for firewall rule and one-way gateway issues.
Frequently asked questions
Is the ot-ics skill free to use?
The skill is published under the MIT licence, which is stated on this page. Check the licence file in the repository for the full terms. Tools such as Wireshark, Nmap, binwalk and Ghidra are installed separately, and commercial products like Claroty or Nozomi are not bundled.
What do I need to install before using it?
The skill files themselves are only instructions. You need to install the analysis tools you use, such as Wireshark with OT dissectors, Nmap NSE, binwalk and Ghidra, and obtain the companion skills firmware-pentest and protocol-reverse, which are not included.
Where do I put the skill folder?
Place the ot-ics folder in ~/.claude/skills/ for personal use or in a project's .claude/skills/ folder for one project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
Can the skill test a live production PLC?
It is not designed for that. The skill forbids writing to PLC coils or registers and high-rate scanning of production OT unless clearly permitted, and it expects a signed scope. Live production or safety-system testing without authorization can cause physical harm and outages.
Can I use it for normal IT or web application testing?
No. The skill states that ordinary web scanning defaults should not be used against OT systems, and it is intended for authorized OT/ICS work only.
How does it compare with commercial OT monitoring platforms?
The skill is a set of instructions for an agent and does not replace a product. It names Claroty and Nozomi only as optional tools a team might use for asset discovery, and those would need their own licences and deployment.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you are a security professional or engineer doing authorized OT/ICS assessments and want a passive-first, safety-focused checklist that an agent follows. Pick something else if you need live production PLC testing, automated active scanning, or general IT and web testing, since the skill deliberately restricts those activities and does not bundle the tools it refers to.
Download the skill