Docker configuration #2
Description
Hi,
I just found this project, which seems to do what I want (run commands inside Docker for guaranteed isolation), except that a great part of the configuration is in a dedicated YAML file. But the JSON list of MCP servers does not show how to use it. I would assume that we should add a --mount option for that file and add the MCP_SHELL_SEC_CONFIG_FILE in the envs. Is it how it is expected to work?
Nevertheless, I am not a fan of having several config files to manage (if I have to do that for each MCP server bringing its own config file, it will become a nightmare). Actually, it is already a general recommendation to use env variables for configuration of services, see point III of the Twelve Factors. And since the content of your YAML does not require such a powerful schema as YAML offers, I would favour following general good practices here.
Would it be possible to replicate the conf in the Docker envs?
I also support #1 which is another step towards a general practice (use remote docker image with local conf), a practice that is bound to generalize for enterprises and other secure tools.