Create IoT architecture diagrams in PlantUML with device, sensor, gateway, edge, and cloud service icons. The iot skill is best for smart home, industrial IoT, fleet telemetry, sensor networks, digital twins, and robotics. Use it for clear IoT vocabulary and iconography, not generic cloud or UML diagrams.
This skill scores 84/100, which means it is a solid listing candidate for directory users. It clearly tells agents when to use it, what output to generate, and which IoT diagram patterns it supports, so users can install with reasonable confidence that it will reduce guesswork for common IoT architecture diagrams.
- Strong triggerability: the description clearly scopes the skill to IoT architecture diagrams and explicitly names best-fit use cases like smart home, IIoT, fleet management, edge computing, and sensor networks.
- Good operational clarity: the SKILL.md includes a quick-start, critical rules, and specific PlantUML/stencil syntax guidance, which helps agents execute the skill correctly.
- Helpful example coverage: multiple concrete example files cover distinct IoT patterns such as device management, digital twins, edge computing, fleet telemetry, robotics, sensor networks, and smart factories.
- No install command or supporting scripts/references are provided, so adoption depends mainly on reading the markdown guidance rather than on executable tooling.
- The repository appears focused on diagram generation only; users looking for broader IoT solution design or general cloud/software modeling should not install it for those tasks.
Overview of iot skill
The iot skill helps you create IoT architecture diagrams in PlantUML with device, sensor, gateway, and cloud service stencil icons. It is best when you need a diagram that explains how physical things connect to edge and cloud systems: smart home, industrial IoT, fleet telemetry, sensor networks, digital twins, and robotics.
This iot skill is not a general diagramming shortcut. It is aimed at readers who need the right IoT vocabulary and iconography, not just boxes and arrows. If your goal is to document data flow, site boundaries, protocol hops, or device lifecycle in a way engineers can scan quickly, this skill fits well. If you only need generic cloud infrastructure or software UML, use a different skill.
What iot is for
Use iot for Diagramming when the diagram must show how devices, gateways, and services relate in the real world: what is deployed on the floor, what runs at the edge, what reaches the cloud, and what gets grouped by site or zone. The strongest value is clarity around IoT-specific components and patterns, not decoration.
When this skill is the right fit
Choose the iot skill if you need diagrams for:
- smart home hubs and device groups
- factory lines, PLCs, sensors, and edge PCs
- vehicle fleets and telemetry pipelines
- LoRaWAN or gateway-aggregated sensor networks
- digital twin and asset-model flows
- robotics or edge ML deployments
What makes it different
The main differentiator is stencil-driven PlantUML output using AWS-style IoT icons such as sensors, gateways, Greengrass, IoT Core, SiteWise, FleetWise, and device management components. That gives you a more credible iot guide for architecture reviews than plain rectangles, especially when the audience already expects service-specific symbols.
How to Use iot skill
Install and load the skill
Use the standard repo install flow for an iot install:
npx skills add markdown-viewer/skills --skill iot
Then open the skill files in this order:
SKILL.mdfor rules and syntaxexamples/*.mdfor patterns you can reuse- any linked stencil reference mentioned in the skill body
Start from the diagram job, not the icon list
A good prompt for this skill should describe the actual architecture outcome, not just “make an IoT diagram.” Include:
- the domain: smart home, factory, fleet, etc.
- the device types
- the edge layer, if any
- the cloud services involved
- the protocol or data path
- the audience, if it changes the level of detail
Example of a strong request:
“Create an iot architecture diagram for a smart factory line with temperature and vibration sensors, a PLC, Greengrass at the edge, IoT Core, SiteWise, and event-based alerting. Show the data path from sensors to edge to cloud and group components by production line and cloud platform.”
Read the examples before writing your own prompt
The example files are the fastest way to learn the skill’s preferred patterns:
examples/smart-home.mdexamples/smart-factory.mdexamples/edge-computing.mdexamples/fleet-telemetry.mdexamples/digital-twin.mdexamples/sensor-network.mdexamples/device-management.mdexamples/robotics.md
These show which stencils pair well together and how the diagrams are usually grouped. That matters more than trying to memorize every icon name.
Use the skill rules that affect output quality
A few rules materially change the result:
- wrap output in
```plantumlor```puml - start with
@startumland end with@enduml - use
left to right directionfor typical device-to-cloud flows - group systems with
rectangleorpackage - prefer directed arrows for flow and dashed arrows for async updates
- do not ask for generic color tuning unless you truly need it
If you already know the target environment, say so. For example, “use MQTT from sensors to gateway” or “show OPC-UA from PLC to edge.” Those details improve the iot usage because they force the diagram to reflect actual integration, not just visual structure.
iot skill FAQ
Is iot only for AWS diagrams?
No. The skill uses AWS-style mxgraph.aws4.* stencils, but the diagrams are still useful for broader IoT architecture. The important part is the IoT visual language and PlantUML workflow, not vendor lock-in.
Can I use this instead of a normal prompt?
Yes, if you want consistent IoT-specific symbols and structure. A plain prompt can describe a diagram, but the iot skill gives you a more repeatable pattern for device, edge, and cloud composition.
Is the iot skill beginner-friendly?
Yes, if you can describe your system in plain English. You do not need to know PlantUML syntax in advance, but you do need to know the devices, flow, and major services you want shown.
When should I not use iot?
Do not use it for generic app architecture, CRUD backends, or standard cloud-only diagrams. The iot skill is most useful when physical devices or edge processing are central to the story.
How to Improve iot skill
Give the skill the real deployment shape
The best iot guide inputs are concrete: how many zones, where the edge sits, and what is local versus cloud-based. “Factory with sensors” is weak. “Two production lines, each with temperature and vibration sensors, one Greengrass gateway per line, and centralized SiteWise analytics” is much better.
Specify the protocol and data path
Output quality improves when you name the path explicitly:
- sensor → gateway → core
- PLC → industrial PC → cloud
- vehicle edge agent → telemetry service → analytics
This helps the iot skill choose better arrows, labels, and grouping. It also prevents a generic diagram that looks plausible but hides the real integration.
Watch for the common failure modes
The most common problems are vague scope, too many services, and missing boundaries. If you do not say what belongs at the edge, the diagram may flatten everything into one layer. If you over-request services, the result can become crowded and harder to read than a simpler iot for Diagramming output.
Iterate by tightening one layer at a time
If the first diagram is too broad, refine it in layers:
- confirm the physical devices
- confirm the edge components
- confirm the cloud services
- confirm the grouping by site, fleet, or product line
That iteration style works especially well for iot install workflows because the skill is strongest when the architecture is already known and you want a clean, service-aware visual translation.