To use a PowerShell script on a self-hosted Linux runner, you must explicitly set the shell to pwsh in your workflow step. This tells GitHub Actions to use PowerShell Core, which is the cross- platform version of PowerShell required for Linux.
Key Requirements for Linux Runners
PowerShell Core (pwsh): Ensure powershell (PowerShell 7+) is installed on your Linux VM. On Linux, the executable name is typically pwsh, not powershell.exe.
File Permissions: If the script is stored in your repository, ensure it has executable permissions.
You can set this in Git using git update-index --chmod=+x scripts/compile-and-package.ps1.
Pathing: Use forward slashes (/) for paths or relative paths starting with ./ to avoid ambiguity on Unix-like systems.
Pre-installed Tools: Since you are using "existing tools on the Linux VM," ensure the runner user has the necessary environment variables (like PATH) to access those tools during the PowerShell session.
Note:
Workflow Configuration
Define your job to run on your self-hosted runner and use the following syntax for the script step:
jobs:
build-and-package:
# Use the label assigned to your self-hosted Linux VM
runs-on: [self-hosted, linux]
steps:
- name: Check out repository code
uses: actions/checkout@v4
- name: Run Compilation and Packaging Script
# 'pwsh' is required for PowerShell Core on Linux
shell: pwsh
run: |
# If the script is in your repo:
./scripts/compile-and-package.ps1
# Or call it with specific parameters
# ./scripts/build.ps1 -Configuration Release
Reference:
https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-powershell

To make a GitHub Actions workflow reusable, you must include the workflow_call keyword under the on key.
This specific event trigger allows the workflow to be called by other "caller" workflows rather than only triggering on standard events like a code push or pull request.
Reusable Workflow Syntax
A basic reusable workflow structure using the workflow_call trigger typically looks like this:
on:
workflow_call:
inputs:
# Optional: Define inputs to be passed from the caller
username:
required: true
type: string
secrets:
# Optional: Define secrets to be passed from the caller
env_token:
required: true
Key Components of workflow_call
inputs: Used to define non-sensitive data (strings, numbers, or booleans) that the caller workflow must or can provide.
secrets: Used to define sensitive data that the caller must pass to the reusable workflow.
outputs: Used to map data from jobs within the reusable workflow so they can be used by the caller workflow.
Reference:
https://www.incredibuild.com/blog/best-practices-to-create-reusable-workflows-on-github-actions

GitHub Actions evaluates each step by the exit status of the process it runs. The runner marks a step successful when the command finishes with an exit status of 0. If the command returns any other code then the step is marked as failed and the job may stop depending on your workflow configuration.

Storing an action in its own repository makes it easier for the GitHub community to discover the action [B], narrows the scope [Not C] of the code base for developers fixing issues and extending the action, and decouples the action's versioning from the versioning of other application code
[D].
Reference:
https://docs.github.com/en/actions/how-tos/create-and-publish-actions/manage-custom-actions