[chis]

Really good idea, I’ll have to try it out. The thing I really want is to have the ability to give a recipe for a new Claude code instance - spin up a docker image with code, data, and a running server and then let Claude work against that.

[smithclay]

Hey, it's not well known but Anthropic actually publishes their own devcontainer feature layer for this (1). Many different CLI wrappers and tools are starting to imbed it (2) but is a nice DIY/open-source way to sandbox.

1. https://github.com/anthropics/devcontainer-features 2. shameless plug, my own open-source CLI for doing this: https://github.com/smithclay/claudetainer

[astral_drama]

I've developed a system that checks out a git repo into a docker compose container, tells claude to read a feature story, then implement it with --dangerously-skip-permissions turned on within the container. It work well enough.

I think it's another workflow, that might suit your needs, but I think the real magic continues to be in the model. Similarly, I'm sure Claudia adds some nice window dressing.

Eventually folks will settle in on some local maxima for interaction and software development with LLMs. Who know what it will look like? It'd be nice if whatever comes next bumps the industry out of the current scrum and sprint style workflows, if only to break folks out of the theatrics of these software personas and rituals.

Here's my WIP: https://github.com/astral-drama/filter

[chrisdemarco]

I’ve used something similar, it’s a good workflow!

[TheTaytay]

I’ve been wanting the same thing. I am currently experimenting with a container that adds what I want. I keep wanting my sandboxed dev machine that is preconfigured with Claude (and almost nothing else)…but then I also want my customized Claude. I saw someone here on HN the other day who just spins up a Docker container and shares their home directory Claude files with it to facilitate the customized Claude container:

https://news.ycombinator.com/item?id=44549802

[adastra22]

You can do this with devcontainer and still work within VSCode.

[vunderba]

Agree would feel far more comfortable if each Claudia project was a separate docker container that was spun up with the appropriate tooling (Node, Python, etc.) particularly with flags like `--dangerously-skip-permissions`.

There is an outstanding feature request around it at least:

https://github.com/getAsterisk/claudia/issues/8