More screenshots
Works with the AI Agents you already use
OT and ICS Security Assessment
Assess plant and utility control networks without endangering the process: written authorization first, passive-first observation, Purdue zoning, findings rated by process impact.
$5
OT and ICS Security Assessment
Example session with this skill installed
Our wastewater plant has never had a security assessment. Corporate IT wants to vulnerability-scan the control network next week. Tell us what to do instead, and what an assessment under this method looks like.
- Read your context and instructions
- Compiled the ot and ics
VERDICT ON THE SCAN — Do not let IT run it. A general-purpose vulnerability scan against a live control network is the exact failure mode this method prevents: fragile controller stacks fault on scan traffic, and no written authorization for that scan exists yet. WEEK 0 — AUTHORIZATION: Two plant networks in scope, levels 1–2 only (paper review + passive observation), no active scanning, no writes. Abort contact: the shift supervisor, named, with stop authority. LEVEL 1, PAPER (DAYS 1–3): Firewall ruleset review shows the corporate VLAN can reach the PLC VLAN on Modbus/TCP. That is a Critical finding from documents alone, zero packets sent. Workstation inventory shows vendor default credentials still set on the engineering laptop. LEVEL 2, PASSIVE (DAYS 4–10): Mirror port on the control switch. Observed: four PLCs and one HMI speaking Modbus/TCP, plus an unexpected SMB flow from the engineering laptop to the historian. No safety-system traffic on this segment — recorded, because it decides what any future active phase may touch. FINDINGS REGISTER (EXTRACT): - F-01 Critical / Production impact: flat path from corporate IT to the treatment PLCs. Compensating control: deny rule plus a jump host with unique credentials, this month, no outage needed. - F-02 High / Production: unauthenticated Modbus — any host on the segment can write process values. Control: segment the PLC VLAN. - F-03 High / Quality: unpatched engineering laptop is the pivot point. Control: isolate now, patch at the next planned outage. NOT TESTED, BY DESIGN: no active probes, no credential tests against production controllers, safety systems untouched. Read-only active checks are proposed for the annual outage window only, under a separate written authorization.
Connects securely to your tools. The creator never sees your data.
About this skill
A defensive assessment method for operational technology and industrial control systems you own or are explicitly authorized to assess. It starts from the constraint that shapes everything in OT: a mistake can have physical consequences, so the standard IT testing playbook does not apply. Before anything is touched, the method requires a written authorization naming the sites and segments in scope, whether any active scanning is permitted (default: no), whether any controller writes are permitted (default: never), the maintenance windows, and an abort contact who can stop the work. The assessment then runs passive-first: offline review of diagrams, firewall rulesets, and PLC project files; passive traffic observation from mirror ports; and only then, if authorized, low-rate read-only checks inside maintenance windows. Findings are mapped onto the Purdue model and rated twice, cyber severity and process impact, with any path down to the control level raising the rating. Recommendations lead with compensating controls (segmentation, jump hosts, passive monitoring) because patching a controller usually waits for a planned outage. Includes an assessment plan template, a pre-fieldwork checklist, a two-axis rating table, a worked water-utility example, and explicit abort conditions.
- It directs and structures the assessment; it does not capture traffic or run tooling against your network — observation happens in your environment with your access. - Ratings are calibrated judgment, not a certification; regulated operators may also need a formal audit against their sector standard. - Depth depends on owner cooperation: without rulesets, diagrams, or a capture point, the assessment stays at the paper level.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
Permissions
No special permissions declared or detected