Prevent the user from accessing or modifying other users’ repositories or folders.

[Rockstarravo]

Currently, users have the ability to access or modify repositories and folders that belong to other users. To ensure proper security and data privacy, we need to implement restrictions that prevent users from reading or writing to repositories or folders that do not belong to them.

1.	Users should only be able to access their own repositories and folders.
2.	Users should not have read or write permissions for other users’ repositories or folders.
3.	Implement necessary access control measures to enforce these restrictions.
4.	Ensure that admin users retain necessary access privileges as required.

Please assist to me on above enforcement

[code-asher]

The only way to achieve this right now is to launch a separate code-server process for each user. You can either use permissions to restrict the file access, or you can use separate containers. You could use something like https://github.com/coder/coder to achieve this.

[Rockstarravo]

This approach may not be feasible, considering we work with over 200 team members. Assigning a separate instance to each user would be a significant undertaking. Are there any alternative solutions we could explore?”

[code-asher]

Not that I am aware of, fundamentally code-server is meant to run as a single-user process.

It would be technically possible to rewrite code-server and VS Code to support multiple users, but I think this would be quite a big undertaking, and it would never be as secure as using containers (not to mention things like port conflicts).

If it helps, there are folks that use coder/coder with well over 200 developers, although I am not sure what their costs look like. Depends on the exact architecture chosen, I imagine.

[benz0li]

Are there any alternative solutions we could explore?

Containerised Juypter, i.e. JupyterHub + b-data's/my JupyterLab docker stacks (which include code-server).
🔬 Check out the reference deployment [using GitHub authentication] at https://demo.jupyter.b-data.ch/.

ℹ️ Jupyter (JupyerHub + JupyterLab) is well established in the Data Science community.

(Test locally: docker run --rm -p 8888:8888 -ti glcr.b-data.ch/jupyterlab/python/base:3.12)