Logic Guard
Overview
Logic Guard is the governance agent for Flowon on Microsoft Dataverse. It reads a project and scans every artifact you have built, logic flows, logic blocks, business processes, schedules, events, event handlers, configurations, localized resources, service connections, and recipes, against a catalog of 143 rules. It returns a governance report: a list of findings, each classified as Critical, Warning, or Info, each with a plain-language explanation of what is wrong and a suggested fix, grouped and counted by section so you can see exactly where the risk sits.
Think of it as a linter for your Flowon configuration. Just as a code linter catches bugs, dead code, and unsafe patterns before they reach production, Logic Guard catches broken transitions, dead outputs, unbounded queries, missing authentication, stale field references, and dozens of other issues across your whole project, automatically and consistently.
It runs as a self-contained container with its own web application, connects through the Flowon MCP server, and is strictly read-only. It reports problems; it never changes your project.
The problem it solves
Quality in a low-code platform is hard to enforce by eye. A project accumulates hundreds of artifacts built by different people over months. Standards drift. A flow ships with a silent error handler. A schedule points at a flow that was renamed. A configuration key is referenced but never defined. A field that a validation depends on gets removed. None of these show up as a red error in the designer, and none of them are caught by a human skim of a large project. They surface later, in production, as the incident nobody can explain.
Logic Guard turns "we should really review this" into an automatic, repeatable gate. Run it before publishing, on a schedule, or as the definition-of-done for every change. The same project always produces the same findings, so quality becomes measurable instead of anecdotal.
What it produces
| Deliverable | What you get |
|---|---|
| Per-artifact findings | Each issue classified Critical, Warning, or Info, with an explanation and a suggested fix |
| Per-section counts | Critical, Warning, and Info totals for every artifact type, with the worst sections ranked |
| Live metadata validation | Stale field and deleted-entity references caught against the real Dataverse schema |
| Cross-artifact findings | Problems that only exist in the relationships between artifacts |
| A customizable ruleset | Every rule can be enabled, disabled, or reclassified per project |
Why counts, not a score
Logic Guard deliberately does not reduce your project to a single "governance score." A score hides the one Critical finding that matters behind ninety cosmetic Info items. Instead it reports Critical, Warning, and Info counts per artifact and per section, and ranks the worst sections, so the report always points you at the highest-severity work first. This keeps the output actionable.
Who it is for
- Reviewers and leads who need a consistent quality gate before a solution is published or promoted.
- Platform teams who want to enforce build standards across many makers.
- Auditors who need evidence that a project meets a defined bar.
- Makers who want fast feedback on their own work before someone else reviews it.
How it fits the platform
Logic Guard reviews everything you build with Logic Composer and Process Orchestrator, reading it through Flowon Logic Assist (the MCP server). It is the quality gate that sits between building and publishing.
Explore Logic Guard
| Page | Contents |
|---|---|
| How It Works | The checker fleet, the cross-artifact pass, and the scan pipeline |
| Rule Catalog | All 143 rules, by section, with severity |
| Customization | Enabling, disabling, and reclassifying rules; strictness and scan context |
| Configuration & API | Per-project settings and the REST surface |
Ready to run it? See the Installation guide.