Clarification on what “preconfigure Copilot's environment” actually means for copilot-setup-steps.yml

[yashmahaja]

Why are you starting this discussion?

Question: Clarification on what “preconfigure Copilot's environment” actually means for copilot-setup-steps.yml

What GitHub Actions topic or product is this about?

Actions Runner

Discussion Details

Hi team!

I’m trying to clearly understand what this line from the Copilot Coding Agent docs means:

Instead, you can preconfigure Copilot's environment before the agent starts by creating a special GitHub Actions workflow file, located at .github/workflows/copilot-setup-steps.yml within your repository.

From testing and community threads, it seems that copilot-setup-steps.yml behaves just like a regular GitHub Actions workflow meaning:

1. It can modify the repository contents (by committing files, generating artifacts, etc.) and Copilot will then see those repository changes when the coding session starts. But the Copilot agent does not appear to use the environment of the setup workflow itself.
   For example, anything done inside the setup workflow’s VM (pip installs, AWS logins, ~/.aws files, Docker images, system-level changes, DNS configuration, etc.) does not carry over into the Copilot agent’s environment.

So my understanding is:

The workflow only prepares the repo, not the actual runtime environment that the Copilot SWE agent will use.
The agent still starts in its own fresh GitHub-hosted runner when the issue/PR session begins.

Could you please confirm whether this interpretation is correct?

Thanks!

[herrksissoumarouane-glitch]

The line in the Copilot Coding Agent docs about “preconfiguring Copilot’s environment via .github/workflows/copilot-setup-steps.yml” can be misleading if taken literally. Here’s what happens:

Purpose of copilot-setup-steps.yml

It’s a GitHub Actions workflow that runs before the Copilot session begins.

Its main role is to prepare the repository: add or modify files, generate artifacts, or set up code in a way that Copilot will see when it starts a coding session.

What it does NOT do

It does not modify the Copilot agent’s runtime VM or environment.

Any system-level changes (like installing Python packages, AWS logins, Docker images, modifying DNS, etc.) do not persist into the agent’s runtime.

The Copilot agent always starts in a fresh GitHub-hosted runner, isolated from whatever happened in the setup workflow.

Effectively:

The setup workflow prepares the repo, not the runtime environment.

The agent’s coding session sees the repo state as it exists after the setup workflow runs, but any VM-level environment changes in that workflow do not carry over.

✅ So your understanding is correct: copilot-setup-steps.yml is about repo preparation, not about preconfiguring the agent’s VM environment.

[yashmahaja]

Thanks for the clarification!!

[JosephKibithe]

You are correct in your interpretation regarding what "preconfigure Copilot's environment" means for copilot-setup-steps.yml in the context of the Copilot Coding Agent.

The copilot-setup-steps.yml workflow can be used to prepare your repository before a Copilot Coding Agent session starts, but it does not modify the runtime environment (such as the VM, installed tools, system-wide configuration, etc.) that the Copilot agent itself will use for coding tasks.

Specifically:

The setup workflow runs as a normal GitHub Actions workflow, and any changes it makes to the repository (committed files, generated artifacts, etc.) are visible to the Copilot agent when it starts its coding session.
Changes that only affect the workflow runner’s VM (like installing packages, setting up credentials, changing system files, etc.) do NOT persist to the Copilot agent’s environment. The Copilot agent runs in its own fresh GitHub-hosted runner with its own ephemeral environment.
The agent will only see modifications that are present in the repo or as artifacts/files checked into version control. It won’t inherit local environment changes, installed CLI tools, or in-memory state from the runner that executed copilot-setup-steps.yml.
So, in summary: The copilot-setup-steps.yml workflow is for preparing the repository, not the live runtime of the Copilot SWE agent. The agent always starts in an isolated environment, and only repo content or version-tracked artifacts are transferred over.

[github-actions[bot]]

🕒 Discussion Activity Reminder 🕒

This Discussion has been labeled as dormant by an automated system for having no activity in the last 60 days. Please consider one the following actions:

1️⃣ Close as Out of Date: If the topic is no longer relevant, close the Discussion as out of date at the bottom of the page.

2️⃣ Provide More Information: Share additional details or context — or let the community know if you've found a solution on your own.

3️⃣ Mark a Reply as Answer: If your question has been answered by a reply, mark the most helpful reply as the solution.

Note: This dormant notification will only apply to Discussions with the Question label. To learn more, see our recent announcement.

Thank you for helping bring this Discussion to a resolution! 💬