🚨 Mission 06: Add Skills

🕵️ CodenameOPERATION SKILL BOOST⭐ Difficulty⏱️ Time 30 min 🧩 ProductsMicrosoft Copilot Studio Microsoft Power Platform🏷️ TagsCustom Skills Agent Instructions🏭 IndustriesIT

NOTE

Skills are unique to agents powered by the GitHub Copilot harness in Copilot Studio. They aren't available to agents powered by the standard or Copilot chat harness.

🎯 Mission Brief

Your Contoso IT Concierge Agent can answer questions, but a reliable helpdesk also needs focused, reusable behavior for specific tasks. The GitHub Copilot harness introduces Skills for this purpose: self-contained capabilities that package a name, description, and task-specific instructions so the harness can activate the right behavior when it's needed.

In this mission, you'll explore this GitHub Copilot harness capability by iteratively refining two skills: one for troubleshooting and another for device guidance. The skills will guide the agent to call SharePoint correctly, recover from query issues, map user-selected options to the correct item, and capture additional requirements.

This work turns a general-purpose agent into a more dependable support experience. Instead of placing every procedure in the agent's global instructions, you'll keep task-specific guidance in focused packages that the harness can activate when relevant. You'll inspect the generated skill files, improve their instructions, and test how small wording changes affect tool calls and responses. The result is behavior that is easier to review, reuse, and refine without overloading the agent's core prompt.

IMPORTANT

In the current Copilot Studio UI, make sure the New experience toggle in the upper-right corner is turned on to use the GitHub Copilot harness authoring surface.

🔎 Objectives

Understand why Skills are unique to agents powered by the GitHub Copilot harness
Understand how skills provide reusable, task-specific behavior without authored topic paths
Review and refine an AI-generated skill to improve instruction quality
Validate the updated skill behavior in Preview, including tool input handling and follow-up requirements

🧠 What is a skill?

Skills are reusable capabilities available to GitHub Copilot harness agents in Copilot Studio. Each skill provides focused guidance for a specific scenario, business process, or task by defining when it should be used, what information to collect, which tools to call, and how the agent should respond.

Skills are part of what distinguishes the GitHub Copilot harness from the other Copilot Studio harnesses:

The standard harness uses topics, prompts, rules, and defined conversation paths for structured behavior.
The Copilot chat harness focuses on extending Microsoft 365 Copilot Chat with enterprise knowledge.
The GitHub Copilot harness can activate modular Skills as it reasons through complex, multi-step work.

Skills provide repeatable behavior for focused tasks inside an agent. They are also reusable and shareable: you can package a skill as a SKILL.md file or ZIP package and add it to other agents powered by the GitHub Copilot harness.

A troubleshooting process defined in a skill is a perfect fit: it's multi-step, judgment-driven, and should behave identically every time.

IMPORTANT

Skills run as part of the GitHub Copilot harness. Building, testing, evaluating, and using these agents can consume Copilot Credits through usage-based billing.

📄 Authoring Skills

Skills use a portable Markdown-based format. A SKILL.md file contains the following sections:

Section	Purpose	What it should contain
Name	The skill's identity	A short, descriptive title that helps the model recognize when the skill is relevant. Example: `IT Troubleshooting Procedure`.
Description	The skill's trigger and intent	A concise summary of what the skill does and when it should be used. This helps the model decide whether to invoke the skill. Example: `Guides employees through diagnose-fix-escalate steps for IT issues such as login, Wi-Fi, and device problems.`
Optional frontmatter	Skill metadata and configuration	Structured YAML metadata such as author, version, tags, examples, dependencies, or other settings used for organization and governance. Not typically used for procedural instructions.
Instructions	The skill's execution logic	Detailed guidance the model follows after selecting the skill. This can include procedure steps, decision logic, tool usage instructions, validation rules, response formats, escalation paths, and success criteria.

When a user submits a request, the GitHub Copilot harness evaluates the available skills and uses each skill's name and description to determine which one is relevant. When the harness activates a skill, its instructions guide the agent through that task. By separating specialized capabilities into reusable skills, builders can create agents that are more accurate, maintainable, and aligned to business requirements.

🧩 Principles for Effective Skill Instructions

1. Start with the Purpose

Begin by explaining what the skill does and when it should be used.

Example:

text

Troubleshooting Procedure

Use this procedure when an employee reports a technical issue and needs assistance diagnosing or resolving the problem.

This provides context before the model starts executing steps.

2. Use Clear Step-by-Step Actions

Break the process into logical, ordered steps.

Example:

text

## Steps

- **Understand the Issue** - Ask the employee to describe the problem if sufficient information has not already been provided.
  - What application or device is affected?
  - What error message are you seeing?
  - When did the issue start?
  - Can the issue be reproduced?

Avoid:

text

Help the employee resolve the problem.

Specific instructions produce more consistent outcomes.

3. Define Decision Points

Tell the model what to do when different conditions occur.

Example:

text

- **Troubleshooting Decision Rules**
  - If an error message is provided, use it to identify relevant troubleshooting steps.
  - If no error message is available, ask follow-up questions to gather more information.
  - If the issue is resolved, confirm the solution and end the procedure.
  - If the issue cannot be resolved after available troubleshooting steps, recommend escalation.

Good instructions define:

If X happens, do Y
Otherwise, do Z

4. Specify Required Inputs

If information is required, state exactly what needs to be collected.

Example:

text

- **Troubleshooting Information** - Before starting diagnostics, collect:
  - Device type
  - Operating system
  - Application or service affected
  - Error message (if available)
  - Steps that reproduce the issue

This prevents the model from skipping important information gathering.

5. Define Tool Usage

When a skill uses actions or connectors, tell the model when and how to use them.

Example:

text

- **Use the `X` tool** - Only use information returned from the tool, do not invent troubleshooting steps
  - Retrieve troubleshooting procedures
  - Find known issues
  - Identify recommended fixes

Be explicit about parameters, field names, and expected inputs.

6. Describe the Expected Output

Specify how results should be presented to the user.

Example:

text

Present troubleshooting guidance using the following format.

| Step | Action |
| --- | --- |
| 1 | Verify network connectivity |
| 2 | Restart the affected application |
| 3 | Confirm whether the issue persists |

Included:

Clear actions
Expected outcomes
Next steps if the action fails

Without output instructions, responses may become inconsistent.

7. Include Validation Rules

Define what constitutes valid information and successful completion.

Example:

text

- Only present troubleshooting steps returned from approved sources.
- Do not recommend fixes that are not documented.
- Verify whether each step resolved the issue before proceeding.
- Do not assume the issue is resolved unless the employee confirms it.

Validation rules help prevent hallucinations and incorrect recommendations.

8. Define Scope and Boundaries

Tell the model what it can and cannot do.

Example:

text

This procedure only supports troubleshooting approved enterprise applications and devices.
Do not:
- Perform account administration tasks
- Reset passwords
- Approve software requests
- Answer unrelated HR or policy questions

Scope boundaries help keep the agent focused and predictable.

9. Include Escalation Paths

Tell the model what to do when it cannot continue.

Example:

text

If the issue remains unresolved after all available troubleshooting steps:
- Recommend escalation to the IT Help Desk.
- Provide a concise summary including:
  - Device or application affected
  - Symptoms reported
  - Troubleshooting steps attempted
  - Results of each step

This ensures a smooth handoff to support teams.

10. Define Success Criteria

Explain when the procedure is considered complete.

Example:

text

The troubleshooting procedure is complete when:

- The issue is resolved and confirmed by the employee, or
- The issue is escalated with a complete support summary

Clear success criteria tell the model when to stop.

TIP

For Copilot Studio skills specifically, the strongest instructions are usually those that answer four questions clearly:

When should this skill be used?
What information must be collected?
What actions or tools should be called?
What should the final response look like?

If those four areas are well-defined, the skill tends to behave much more consistently.

🧪 Lab 06 - Add the skills

In this lab, you will extend the agent you created in Mission 04 - Build an agent with the GitHub Copilot harness by creating a series of custom skills.

✨ Use case

As an employee

I want to get quick and accurate IT support for common issues and device requests

So that I can stay productive and resolve technical issues faster

Prerequisites

Contoso IT Concierge - the agent created in Mission 04 - Build an agent with the GitHub Copilot harness.

6.1 Create a device request skill and test

The first skill we're going to create is one to handle device requests. We'll take an iterative approach showing how to test and refine a skill.

Download the skill package using the button below.
Download device-guidance-v1-0-1.zip and extract it.
In the Build tab of your agent, select the Add + button next to the Skills section.
Make sure the Upload a skill tab is selected. Then drag and drop the extracted SKILL.md file, or locate the file and upload it.
The skill will then be uploaded. Select the skill in the right-hand panel to review the instructions.
- The description references the name of the tool
- The instructions outline steps including
  - the purpose and trigger
  - clear step-by-step actions, including aligning the selected device option to the corresponding SharePoint Item ID
  - tool usage where the Get Employee Assets tool is referenced
  - expected output that outlines the format of the response
  - validation rules
  - exception and escalation paths
  - success criteria
Select the X icon when you're done reviewing to close the skill.
Reminder of the skill structure
A skill consists of four main components.
1. The Name identifies the skill.
2. The Description tells the model when the skill is applicable.
3. The Optional Frontmatter stores metadata and configuration information.
4. The Instructions contain the operational workflow the model follows to complete the task.
Together, these components help the model determine when to use the skill and how to execute it correctly.
Next, update the agent instructions to explicitly reference the skill for device requests. Under the For device requests: heading, delete the existing bullet points.
Copy and paste the following text as a new bullet point under the For device requests: heading.
text
```
Refer to the `device-guidance-v1-0-1` skill
```
Save the agent.
Select Preview at the top center of the agent and select New chat to test the updated skill.
Copy and paste the following text and submit it to the agent.
text
```
I need a laptop
```
The agent invokes the skill according to the updated instructions.
You may encounter an error if the model guesses a column name that doesn't exist in the SharePoint list.
The following are examples of errors you may encounter.
Example 1:
Example 2:
However, what happens next is that the model uses reasoning and dynamic planning to determine the next appropriate action when it encounters an issue.
In this step, the model applies reasoning and uses the correct SharePoint internal column name, field_4 (the Asset Type column), to filter for the device type Laptop.
The agent then displays the available laptops from the SharePoint list.
Regardless of whether you encountered an error, refine the skill inputs. As you learned in the Define Tool Usage principle, effective skill instructions should explicitly define parameters, field names, and expected inputs.
Download the skill package using the button below.
Download device-guidance-v1-0-2.zip and extract it.
In the Build tab, select the device-guidance-v1-0-1 skill, select the ellipsis, select Replace
Upload the extracted SKILL.md file into the agent.
The updated skill now explicitly maps SharePoint display names to their internal column names. This prevents the model from guessing field names and ensures OData filters use valid columns, such as field_4 for the Asset Type column. It also includes an example showing how to construct a filter correctly.
How to find the SharePoint column schema name
1. In your SharePoint list, select the Settings icon, and then select List settings.
2. Under Columns, select the column whose internal name you want to find, such as Asset Type.
3. In the edit-column page URL, locate the Field query parameter. For example:
  text
```
.../FldEdit.aspx?List={LIST-ID}&Field=field%5F4
```
The value of Field is the column's internal schema name. SharePoint URL-encodes _ as %5F, so Field=field%5F4 corresponds to the internal name field_4. Use this internal name when constructing the OData filter query.
Next, update the agent instructions to reference the updated skill. Replace the bullet point under the For device requests: heading with the following text.
text
```
Refer to the `device-guidance-v1-0-2` skill
```
Save the agent.
Test the updated skill by navigating to Preview.
Copy and paste the following text and submit it to the agent.
text
```
I need a laptop
```
The agent invokes the skill. Expand the skill details and confirm that device-guidance-v1-0-2 is referenced.
The error should no longer appear because the updated skill instructions guide the model to use the correct tool inputs when constructing the filter query.
In this test, respond to the agent by providing the device option you want to proceed with.
Copy and paste the following text and submit it to the agent.
text
```
A
```
Following the skill instructions, the agent summarizes the device selection in its response.
We're not done refining the skill yet. If a user wants to provide an additional comment as part of the request, that behavior also needs to be captured in the skill instructions. Let's update the skill again.
Download the skill package using the button below.
Download device-guidance-v1-0-3.zip and extract it.
Repeat the same steps to replace the skill:
- In Build, select the device-guidance skill
- Select the ellipsis icon and select Replace
- Upload the extracted SKILL.md file into the agent
The updated skill now instructs the agent to ask about additional device requirements after the user selects an option. It captures specifications such as RAM or storage capacity and includes them in the device request. In the next mission, you'll pass this information into a workflow for the Contoso IT Concierge agent.
Next, update the agent instructions to reference the updated skill. Replace the bullet point under the For device requests: heading with the following text.
text
```
Refer to the `device-guidance-v1-0-3` skill
```
Save the agent.
Test the updated skill by navigating to Preview.
Copy and paste the following text and submit it to the agent.
text
```
I need a laptop
```
The agent invokes the skill. Expand the skill details and confirm that device-guidance-v1-0-3 is referenced.
In the response, copy and paste the following text and submit it to the agent.
text
```
A
```
The agent responds by next asking the user for any additional requirements they may have for their device.
Copy and paste the following text and submit it to the agent.
text
```
16GB of RAM
```
The agent responds by summarizing the selected device and the additional requirement.

Notice that the agent can't complete the device request because it hasn't been configured to submit requests. In the next mission, you'll add a workflow that enables the agent to proceed with submitting the device request.

6.2 Create a troubleshooting skill and test

Now we'll work on a separate skill that specializes in troubleshooting issues.

To use the troubleshooting skill, download the skill package using the button below.
Download it-troubleshooting-procedure.zip and extract it.
In the Build tab of your agent, select the Add + button next to the Skills section.
Make sure the Upload a skill tab is selected. Then drag and drop the extracted SKILL.md file, or locate the file and upload it.
The skill will then be uploaded. Select the skill in the right-hand panel to review the instructions.
Select the X icon when you're done reviewing to close the skill.
Reminder of the skill structure
A skill consists of four main components.
1. The Name identifies the skill.
2. The Description tells the model when the skill is applicable.
3. The Optional Frontmatter stores metadata and configuration information.
4. The Instructions contain the operational workflow the model follows to complete the task.
Together, these components help the model determine when to use the skill and how to execute it correctly.
Next, update the agent instructions to reference the updated skill. Replace the bullet point under the For troubleshooting: heading with the following text.
text
```
Refer to the `troubleshooting-procedure` skill
```
Save the agent.
Select Preview at the top center of the agent to test the updated skill.
Copy and paste the following text and submit it to the agent.
text
```
I'm not getting new emails anymore
```
The agent invokes the skill. Expand the skill details and confirm that troubleshooting-procedure skill is referenced.
The agent asks a follow-up question.

In this test, respond with the following:

text

Happening on both outlook on windows and outlook mobile on my ios device

Reply to the troubleshooting question

Following the skill instructions, check that it responds with a numbered list of troubleshooting steps to try and asks another follow-up question. Reply with the following:
text
```
The Windows Outlook status bar is showing working offline
```
Check that it replies with a suggested fix and reply with the following:
text
```
That fixed it, I'm receiving email now!
```
The agent should then close out with a message confirming the fix.

✅ Mission Complete

Mission accomplished, Recruit! You used a capability unique to the GitHub Copilot harness and iteratively refined a reusable skill so the Contoso IT Concierge agent can reliably map tool inputs, recover from OData filter errors, match user-selected options to SharePoint item IDs, and capture additional device requirements.

You can now:

✅ Create a reusable agent skill: Package focused instructions and supporting files for a defined task.

✅ Guide tool usage: Map skill inputs to the SharePoint tool and handle OData filter errors.

✅ Validate user selections: Match selected device options to SharePoint item IDs and collect additional requirements.

✅ Define reliable outcomes: Add validation rules, escalation paths, and success criteria that improve consistency.

⏭️ Move to Automate with Workflows to get started.

📚 Tactical Resources

🔗 Skills in Copilot Studio

🔗 Write effective instructions

Recruit

Standard harness

GitHub Copilot harness

Operative

Special Ops

Cowork Collective

🚨 Mission 06: Add Skills

🎯 Mission Brief

🔎 Objectives

🧠 What is a skill?

📄 Authoring Skills

🧩 Principles for Effective Skill Instructions

1. Start with the Purpose

2. Use Clear Step-by-Step Actions

3. Define Decision Points

4. Specify Required Inputs

5. Define Tool Usage

6. Describe the Expected Output

7. Include Validation Rules

8. Define Scope and Boundaries

9. Include Escalation Paths

10. Define Success Criteria

🧪 Lab 06 - Add the skills

✨ Use case

Prerequisites

6.1 Create a device request skill and test

6.2 Create a troubleshooting skill and test

✅ Mission Complete

📚 Tactical Resources

Standard harness

GitHub Copilot harness

🚨 Mission 06: Add Skills ​

🎯 Mission Brief ​

🔎 Objectives ​

🧠 What is a skill? ​

📄 Authoring Skills ​

🧩 Principles for Effective Skill Instructions ​

1. Start with the Purpose ​

2. Use Clear Step-by-Step Actions ​

3. Define Decision Points ​

4. Specify Required Inputs ​

5. Define Tool Usage ​

6. Describe the Expected Output ​

7. Include Validation Rules ​

8. Define Scope and Boundaries ​

9. Include Escalation Paths ​

10. Define Success Criteria ​

🧪 Lab 06 - Add the skills ​

✨ Use case ​

Prerequisites ​

6.1 Create a device request skill and test ​

6.2 Create a troubleshooting skill and test ​

✅ Mission Complete ​

📚 Tactical Resources ​

🚨 Mission 06: Add Skills

🎯 Mission Brief

🔎 Objectives

🧠 What is a skill?

📄 Authoring Skills

🧩 Principles for Effective Skill Instructions

1. Start with the Purpose

2. Use Clear Step-by-Step Actions

3. Define Decision Points

4. Specify Required Inputs

5. Define Tool Usage

6. Describe the Expected Output

7. Include Validation Rules

8. Define Scope and Boundaries

9. Include Escalation Paths

10. Define Success Criteria

🧪 Lab 06 - Add the skills

✨ Use case

Prerequisites

6.1 Create a device request skill and test

6.2 Create a troubleshooting skill and test

✅ Mission Complete

📚 Tactical Resources