Guide · Agent skill
How to use the Microsoft 365 Declarative Agents Kit agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The declarative-agents skill is a set of instructions that teaches your AI agent how to help you build Microsoft 365 Copilot declarative agents. A declarative agent is a custom Copilot assistant defined by a manifest rather than custom code. This guide explains what the skill does, how to install it, how to ask for it, and what you need before you start.
Microsoft 365 Declarative Agents Kitgithub/awesome-copilot · MIT Download the skill Quick start
- 1Download and unzip. Download the ZIP from this page and unzip the declarative-agents folder.
- 2Place the skill. Move the folder into ~/.claude/skills/ for personal use, or into .claude/skills/ in your project.
- 3Install the toolkit. Install the Microsoft 365 Agents Toolkit VS Code extension, teamsdevapp.ms-teams-vscode-extension, if you want the integrated workflow.
- 4Describe your goal. Ask in plain language, for example to create, review or migrate a declarative agent.
- 5Choose a workflow. Pick basic, advanced or validation, or let the agent recommend one based on your situation.
- 6Test locally. Use Agents Playground to test the agent before deploying it to your Microsoft 365 tenant.
What is an agent skill and what does this one do?
An agent skill is a folder containing a SKILL.md file that gives your AI agent extra instructions for a specific task. This skill, named declarative-agents, is a development kit for Microsoft 365 Copilot declarative agents. When it is active, the agent acts as a guide for planning, writing, validating and migrating these agent definitions.
The skill offers three workflows. The basic workflow covers agent planning, capability selection, a JSON manifest, a TypeSpec alternative, and local testing with Agents Playground. The advanced workflow covers enterprise design, including localization, production deployment and monitoring. The validation workflow checks an existing agent against the v1.5 schema limits: name (100 characters), description (1000) and instructions (8000).
- Basic: new developers and quick prototypes
- Advanced: enterprise design, deployment and monitoring
- Validation: existing agents, troubleshooting and optimization
How do you install the declarative-agents skill?
Download the ZIP file for the skill from this page and unzip it. You can install the skill in one of two ways.
For a personal install, unzip the folder into ~/.claude/skills/. For a project install, unzip it into the .claude/skills/ folder inside your project. Alternatively, upload the skill in the Claude apps under Settings → Capabilities → Skills.
- Personal: ~/.claude/skills/
- Project: .claude/skills/ inside your project folder
- Claude apps: Settings → Capabilities → Skills
What do you need before you start?
The skill is only useful inside the Microsoft 365 ecosystem. To deploy an agent you need Microsoft 365 Copilot access and a Microsoft 365 tenant. For the integrated workflow, install the Microsoft 365 Agents Toolkit VS Code extension, whose ID is teamsdevapp.ms-teams-vscode-extension.
Agents Playground is used for local testing. If you write definitions in TypeSpec, you also need TypeSpec tooling.
- Microsoft 365 Copilot access and a Microsoft 365 tenant (for deployment)
- Microsoft 365 Agents Toolkit extension: teamsdevapp.ms-teams-vscode-extension
- Agents Playground for local testing
- TypeSpec tooling, if you write definitions in TypeSpec
How do you trigger the skill with natural requests?
You do not need a special command. Describe what you want in plain language and mention the agent type. The skill is written to take a request such as the examples below and choose the matching workflow.
If you are not sure where to start, ask the agent which workflow fits your situation, as the skill ends by asking which workflow you would like to start with.
- "Help me create a basic declarative agent for answering HR policy questions."
- "Review my existing declarative agent manifest and check it against the v1.5 limits."
- "Convert my agent's JSON manifest to TypeSpec."
- "Plan a multi-language declarative agent for deployment to production."
What limits and capabilities does the skill check?
The skill lists up to 11 capabilities, of which you can choose up to 5: WebSearch, OneDriveAndSharePoint, GraphConnectors, MicrosoftGraph, TeamsAndOutlook, PowerPlatform, BusinessDataProcessing, WordAndExcel, CopilotForMicrosoft365, EnterpriseApplications and CustomConnectors. It also lists array limits of up to 4 conversation starters and up to 5 capabilities.
The validation workflow uses these figures, so check them against current Microsoft documentation before you rely on them.
- Name: 100 characters
- Description: 1000 characters
- Instructions: 8000 characters
- Conversation starters: maximum 4
- Capabilities: maximum 5
What about safety, scripts and network use?
The skill does not run scripts and does not use the network, according to its listing. It works through instructions the agent follows, so you still control what the agent executes and where files are saved.
Because the file gives no source links, verify schema details and any Microsoft guidance yourself before deploying an agent to users in your tenant.
- Runs scripts: no
- Uses network: no
- The skill includes no source links for its schema rules
What people use it for
- Building a simple prototype declarative agent for a small team
- Checking an existing manifest against the v1.5 character and array limits
- Converting a JSON manifest to TypeSpec definitions
- Planning localization and deployment for an enterprise agent
Frequently asked questions
How do I install the declarative-agents skill?
Download the ZIP from this page and unzip it. For a personal install, place the folder in ~/.claude/skills/. For a single project, place it in .claude/skills/ inside that project, or upload it in the Claude apps under Settings → Capabilities → Skills.
What do I need to use it?
You need Microsoft 365 Copilot access and a Microsoft 365 tenant for deployment. The integrated workflow uses the Microsoft 365 Agents Toolkit VS Code extension, and local testing uses Agents Playground. TypeSpec tooling is needed only if you write definitions in TypeSpec.
Is the declarative-agents skill free?
The skill is released under the MIT licence. Microsoft 365 Copilot itself requires access through your Microsoft 365 subscription, which this page does not cover.
Does the skill run scripts or use the internet?
According to its listing, the skill runs no scripts and uses no network. It provides instructions that the agent follows.
What are the character limits for a declarative agent?
The skill lists name at 100 characters, description at 1000 and instructions at 8000, based on the v1.5 schema. Check these against current Microsoft documentation, as the rules may have changed.
Is there an alternative to this skill for building declarative agents?
The skill itself points to the Microsoft 365 Agents Toolkit for VS Code and to TypeSpec as ways to write definitions. You can also write the JSON manifest by hand, though this file does not compare other tools in detail.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits developers new to Microsoft 365 Copilot who want guided help building a declarative agent, and teams checking an existing manifest against schema limits. It is not useful for people outside Microsoft 365, and anyone relying on its v1.5 figures should confirm them against current Microsoft documentation first, since the file contains no source links.
Download the skill