How to use the Smart City IoT Solution Planner agent skill

What the skill makes the agent do

When the skill is active, the agent follows a structured process rather than answering freely. It first checks Azure IoT Edge documentation before making edge computing decisions, then asks about scope, scale, latency, privacy rules and existing systems that must be integrated.

From those answers it produces a set of planning artifacts: a solution summary, a reference architecture, a security and governance checklist, a cost and scaling strategy, and a phased implementation backlog. It is instructed to label assumptions clearly and to separate them from confirmed facts.

Install the skill from the ZIP or the Claude apps

Download the ZIP for this skill from this page and unzip it. To make the skill available to Claude Code for all your projects, place the unzipped folder in ~/.claude/skills/. To make it available only inside one project, place it in that project's .claude/skills/ folder.

If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.

• Personal use across projects: ~/.claude/skills/
• Single project: the project's .claude/skills/ folder
• Claude apps: Settings, then Capabilities, then Skills

Trigger the skill with natural requests

You do not call the skill by name. Describe what you want in plain language that matches the skill's purpose. The skill's own examples include requests about building an IoT solution on Azure, Smart City architecture for traffic, lighting or waste, connecting devices, analytics and alerts, and needing a roadmap and backlog for an urban platform.

The more context you give in the first message, the fewer questions the agent will need to ask. Useful details include the city domain, approximate number of devices, how often they send data, how long data must be kept, and which existing systems must connect.

• I want to build an IoT solution on Azure for street lighting across three districts.
• Smart City architecture for traffic signals and parking sensors, with a pilot in one district.
• How do I connect waste bin sensors, analytics and alerts on Azure?
• I need a roadmap and backlog for an urban air quality platform.

What you need and safety notes

The files do not list any required software or API keys. The skill works best with internet access, because the agent is asked to read the Azure IoT Edge documentation at learn.microsoft.com. If the documentation cannot be consulted, the agent is instructed to say so and continue with clearly marked assumptions.

The skill runs no scripts, but it does use the network. It can also use optional runtime Azure skills and local repository skills if your host environment provides them. The skill is planning-focused: it is written to validate prerequisites before any deployment and does not deploy or change live Azure resources.

• Runs scripts: no
• Uses network: yes, to read Azure IoT Edge documentation
• Optional: runtime Azure skills and local repository skills, if available in your host
• Not a deployment tool: review all output with your own engineers before acting on it

Read and use the planning output

The skill uses a template in references/smart-city-solution-template.md to keep outputs consistent across projects. Its response structure covers context and objectives, proposed architecture, technology decisions and trade-offs, security, operations and cost controls, a phased implementation plan, and risks and open questions.

Treat the output as a draft for review. Its guidance avoids single-region production for critical city workloads and asks for named operational ownership, such as who handles incidents, SLAs and change windows. Cost figures are estimates; check them with pricing tools and your own quotes before you commit budget.

• Phase 1 is a pilot district or a single use case
• Phase 2 adds multi-domain integration
• Phase 3 covers city-scale rollout and optimization
• Each phase lists exit criteria, dependencies, risks, mitigations and KPIs

Tips for better results

Name the city domain and the constraints early, especially privacy rules and latency needs, because the skill asks about them before it designs anything. If you have an existing system such as SCADA, GIS or a ticketing tool, list it so integration is planned rather than assumed.

Ask for a follow-up pass on one section at a time, for example the security checklist or the phase 1 backlog. Smaller requests usually give more specific answers than one broad request.

• Give scale numbers, even rough ones
• List existing systems to integrate
• Ask for one artifact at a time when you need depth