01
Attackers go after the systems that run the business
Warehouse, ERP and transport systems hold your orders, your stock and your customers' data. When one of them stops, the operation stops with it.
That makes them a target. Ransomware and stolen logins are an everyday risk for business systems, and many of those systems were set up years before that was true. They were configured to get the operation running, by people whose job was to ship orders, and the settings chosen then are often still in place.
The weak points are rarely dramatic. A default administrator login that still works. A Java runtime that missed two years of patches. A super-user role that was handed out for one busy week and never taken back. Each one is small, each one is easy to fix, and each one is exactly what an attacker looks for first.
The usual answer is an audit. Someone reads the settings once a year and writes a report. The report is accurate on the day it is written and out of date as soon as a setting changes. Between audits, nobody can say with confidence what state the system is in.
02
What a standard scanner cannot see
Most organisations already run a security scanner, and they should keep running it. A scanner tests the server, the network and the code for known technical weaknesses, and it does that well.
What it does not know is the application. It cannot tell which accounts in the warehouse system have not been used in months, which users hold super-user roles, or that the application still accepts the login it was installed with. Those facts live inside the application, in its own tables and settings, in a form only that application understands.
A security agent starts where the scanner stops. It reads the application through the application's own commands, and checks what it finds against rules written for that system. Think of it as a light, application-aware hardening check.
| Compared on | A security agentInside the application | A standard security scannerOutside the application |
|---|---|---|
| Where it looks | Inside the running application | The operating system, the network and the firewall |
| What it knows | The application's own logins, roles, sessions and components | Known technical weaknesses, the same for every customer |
| Typical finding | A default login that still works, or a super-user role nobody needs | An open port or an unpatched operating system |
| What you get | One grade, findings by severity, each tied to framework controls | A list of technical issues |
| How they fit | Add it. It covers what the scanner cannot see. | Keep it. It covers what a security agent does not. |
The scanner tells you whether the door is locked. The agent tells you who has been given a key.
The two answer different questions, so the choice is not one or the other. ProbeWave is not a replacement for a vulnerability scanner, network scanning or a penetration test. It is the part of the assessment those tools leave to a person with a checklist.
03
Meet ProbeWave
ProbeWave is the security assessment agent we built for Blue Yonder WMS. You choose a warehouse connection, run a full assessment or only the checks you want, and it tells you where that instance stands.
Its checks are read-only. They look for known weaknesses, risky settings and exposures, and they never exploit what they find or change any data.
Each check passes, raises a warning or fails. The instance gets a grade from A to F, set by its most serious finding, and the findings are listed with the most serious first, so the first thing you read is the first thing to fix. Each finding carries its severity, a CVSS score, what it could lead to and the steps to fix it.
Every finding is also tied to the controls it affects in CIS Controls v8, NIST CSF 2.0, PCI DSS 4.0 and ISO 27001:2022. When an auditor asks how a control is met, the answer is already attached to the evidence.
The assessment covers four areas.
Default logins that still work, and accounts that are active but have not been used in more than 30 days.
Users with super-user flags or super-user roles, so excess access is found and can be taken back.
Session keys that could be forged, plain HTTP where encrypted connections are expected, and commands that could be run without a login.
The Java runtime and the libraries the application ships with, checked for versions with known vulnerabilities.
The result appears as a summary card in the chat, and the findings are saved as tables you can query. The Report agent can turn them into a PDF report, or compare several runs to show how the instance is improving.
Because the checks are run by an agent and not by hand, the assessment can be repeated whenever you need it. The grade then describes the system as it is today, and not as it was at the last audit.
04
Audit rules become tools. An agent works from them
There is nothing hidden in how ProbeWave is made, and that is the point. It is built from the same parts as every other agent in Smart Forge.
An audit rule is a sentence, such as "no default account stays active". We write that rule as a tool: a small, reviewed piece of work that reads the relevant facts from the system and reports whether the rule holds. The tool also carries the controls the rule relates to.
The agent is then configured to work from those tools. It runs them, gathers the results, grades the system and explains each finding in plain language. The checking is done by the tools, which behave the same way on every run. The model's part is to organise and explain.
One audit rule, from sentence to finding
- Done by the agent
- Written once by your team and ours
- Your audit rules become toolsEach rule is written as a tool that checks it on the system.
- A new agent works from those toolsIt runs the checks, grades the system and lists what it found, most serious first.
- Findings are tied to your frameworksCIS, NIST, PCI DSS and ISO 27001, or your own internal policy.
- It runs whenever you need itOn demand, with every run saved, so the grade reflects the system as it is now.
05
ProbeWave is one example. We can build much more
ProbeWave is a technical baseline for one system. The method behind it does not belong to warehouses, or to security.
Anything you can state as a rule and check against a system can become a tool, and any set of tools can be given to an agent. So the useful question is not whether ProbeWave fits your system. It is which rules you would like checked, and on what.
Segregation of duties, approval limits, password policy for business accounts. ProbeWave does not check these today. An agent built with you can.
An ERP, a transport system or a database can be held to the same frameworks as the warehouse, so one standard applies everywhere.
The checklist your auditors already use becomes a set of tools, and the agent reports against your wording and your controls.
The same method fits any rule you can state: data quality, set-up drift between sites, or readiness before a go-live.
We build these with you. You bring the rules and the knowledge of how your operation works, and we bring the platform and the experience of having done it for a warehouse system.
The usual way, and the Smart Forge way
06
What to know before you rely on it
Five things are worth knowing, so the grade is never read for more than it says.
- It does not replace your security scannerKeep your vulnerability and network scanning and your penetration tests. A security agent adds the view from inside the application.
- A good grade is not a guaranteeThe agent checks the rules it has been given. A risk nobody wrote a rule for is not in the grade, and people, processes and business controls are outside it.
- A known vulnerability is not proof of a way inA version with a published vulnerability is flagged, and the finding explains why the server may or may not be exposed.
- It reads, and people decideIt never exploits a weakness or changes data. Your team decides what to change, and when.
- ProbeWave is for Blue Yonder WMSThat agent exists today. A security agent for another system is built with you, for that system.