microsoft/project-osmos

microsoft/project-osmos

Description: Project Osmos is the copilot/codex/claude code plugin to interact with the Osmos Data Engineer in Microsoft Fabric to solve complex data engineering tasks

Language: Python

License: NOASSERTION

Stars: 13

Forks: 5

Open issues: 2

Created: 2026-06-16T21:55:20Z

Pushed: 2026-09-18T18:02:58Z

Default branch: main

Fork: no

Archived: no

README:

Project Osmos for Microsoft Fabric

Install Project Osmos in GitHub Copilot CLI, Codex, or Claude Code and use natural language to run data engineering tasks in Microsoft Fabric.

What Project Osmos does

Project Osmos is an AI-powered data engineering workflow for Fabric. Describe the outcome you want, and the agent can:

• Explore data in your Lakehouse.
• Create and test Fabric notebooks.
• Clean, join, aggregate, and validate data.
• Create or update Delta tables and workspace artifacts.
• Run longer workflows while you monitor progress from Fabric task details.

The plugin on your machine starts and monitors the task. Agent reasoning, Spark execution, and OneLake data access run inside Microsoft Fabric.

Before you start

You need:

| Requirement | Details | | --- | --- | | Project Osmos access | Project Osmos must be enabled for your account. | | Fabric workspace | The workspace must be assigned to Fabric capacity and contain a Lakehouse. | | Workspace permissions | Contributor or higher on the target workspace. | | Fabric Copilot setting | User can use Copilot and other features powered by Azure OpenAI must be enabled for the tenant or workspace. | | Azure CLI | Install Azure CLI and sign in with an identity that can access the workspace. Guest users may optionally supply the workspace's resource tenant ID. | | Python 3.11+ | Use an existing compatible interpreter. Project Osmos helpers use only the standard library and do not install packages or modify local environments. | | AI coding client | Install GitHub Copilot CLI, Codex, or Claude Code. |

Install

Use the commands for your preferred client.

GitHub Copilot CLI

copilot plugin marketplace add microsoft/project-osmos
copilot plugin install project-osmos@project-osmos

Codex

codex plugin marketplace add microsoft/project-osmos
codex plugin add project-osmos@project-osmos

Claude Code

claude plugin marketplace add microsoft/project-osmos
claude plugin install project-osmos@project-osmos

Restart the client after the first installation. In Claude Code, you can instead run /reload-plugins.

Create your first task

1. Sign in to Azure

Sign in with the Azure CLI:

az login --allow-no-subscriptions

Verify Power BI token acquisition without printing the token:

az account get-access-token \
--resource https://analysis.windows.net/powerbi/api \
--output none

Project Osmos uses this current session by default. If a guest or cross-tenant session cannot access the workspace, provide the Microsoft Entra tenant ID that owns the workspace when prompted.

2. Start your client

| Client | Command | | --- | --- | | GitHub Copilot CLI | copilot | | Codex | codex | | Claude Code | claude |

3. Ask for Project Osmos

Use a prompt that names Project Osmos and describes the goal:

Use Project Osmos to transform data in my Fabric lakehouse.

The skill asks for:

1. Which Fabric workspace and Lakehouse to use. It can reuse Microsoft Fabric page context, resolve names with Microsoft Fabric Skills, or parse a Lakehouse browser URL. 2. Your complete data engineering instruction. 3. Any optional constraints or additional context. 4. Review of the recommended operating settings before the run starts.

The skill asks for a resource tenant ID only if the current Azure CLI session cannot access the workspace.

4. Describe the outcome

Be specific about source data, transformations, outputs, and validation. For example:

Load Orders and Customers from my lakehouse, remove Orders rows with missing
customer IDs, join the tables, calculate monthly revenue by customer segment,
write the result as a Delta table, and validate the source and output row counts.

Another example:

Create a sales sample CSV in a workspace folder, create a matching Delta table,
and build and test a notebook that ingests the CSV into the table.

5. Monitor the run

Project Osmos runs in Fabric and complex tasks can take up to a day. After each task starts, the plugin prints and opens its Fabric task details link. Continue sending follow-ups and control requests through your coding client. If the browser cannot open automatically, use the printed Fabric URL; the remote task and local recovery setup continue.

A headless local poller handles token refresh, bounded automatic recovery, clarification-loop detection, and audit capture. It does not need any browser page open. Task state and audit records remain under:

./.dataprojects//

Closing the browser does not stop the poller. Shutting down your machine or stopping the poller stops local monitoring and automatic recovery, but does not cancel the Fabric task. When you return, reopen your client and ask it to resume the existing Project Osmos task. Include the task ID if the client cannot identify the previous run.

Update the plugin

Copilot CLI and Claude Code include a startup update check. Codex updates configured Git marketplaces in the background. You can also update manually.

GitHub Copilot CLI

copilot plugin marketplace update project-osmos
copilot plugin update project-osmos@project-osmos

Codex

codex plugin marketplace upgrade project-osmos

Claude Code

claude plugin marketplace update project-osmos
claude plugin update project-osmos@project-osmos

Restart the client after an update so the new skill content is loaded.

Troubleshooting

| Issue | What to do | | --- | --- | | Project Osmos is not available after installation | Restart the client, then run the install commands again and check for an installation error. | | Azure authentication fails | Run az login --allow-no-subscriptions and retry. For a guest or cross-tenant workspace, sign in with az login --tenant --allow-no-subscriptions. | | Workspace or capacity lookup fails | Confirm the Lakehouse URL is correct, the workspace has Fabric capacity, and your Azure identity can access it. | | Lakehouse lookup or task creation fails | Confirm the workspace...