SkillFoxx

Devsy CLI

CLI and app for workspaces: each AI agent gets its own container from a shared devcontainer.json, with an MCP server and a skill

  • 137 ★ on the repository
  • Last commit
  • MPL-2.0
CLI

Broad access

Why this level

  • Creates and deletes container workspaces and executes commands in them through workspace_exec on Docker, Kubernetes or the cloud
  • Passes Git and Docker credentials into the container, while network and volume access is set by provider settings
All reasons and checks

devsy-org/devsyAuthor:devsy-org

Install

Manual install

devsy mcp serve

Starts the stdio MCP server for an agent. Devsy itself and providers are installed per the docs at devsy.sh.

Access and checks

How we review

Broad accessThe tool can change data, run commands or handle secrets.

  • Creates and deletes container workspaces and executes commands in them through workspace_exec on Docker, Kubernetes or the cloud
  • Passes Git and Docker credentials into the container, while network and volume access is set by provider settings

This is third-party code. Review the repository files before installing.

What it does

Devsy gives teams and their AI agents separate workspaces built from one standard definition. Workspaces run on Docker, Kubernetes, cloud providers and SSH remote hosts. Each agent starts in its own container built from the same devcontainer.json the team uses. Container access to resources depends on the provider, mounts, network and credentials. It uses the open DevContainer standard, so workspaces are portable across backends. There is auto-shutdown of idle workspaces, prebuilds, Git and Docker credential sync and support for VS Code and JetBrains. The desktop app manages workspaces and the command line serves automation and platform integration. For agents the repository holds a devsy skill and an MCP server: devsy mcp serve provides the workspace_list, workspace_status, workspace_create, workspace_start, workspace_stop, workspace_delete and workspace_exec tools plus provider management. The skill teaches the agent to choose between MCP and the command line, reuse existing workspaces and not replace Devsy operations with direct edits of Docker, Kubernetes or the cloud.

Who it is for. Teams that run several coding agents in parallel and want to isolate them in containers from a shared environment definition.

Good fit when

  • Several agents work on one repository and each needs its own isolated environment
  • An environment has to start locally in Docker and from the same definition on Kubernetes or in the cloud
  • An agent should create, stop and delete workspaces itself through MCP
  • Idle agent workspaces should shut down automatically
  • Git and Docker credentials should reach the container without manual copying

Not a fit when

  • One developer is fine with a regular terminal: the container layer is not needed
  • The project is not described through DevContainer: Devsy builds the environment from that definition
  • The container should be closed from the network and volumes by itself: provider and mount settings define access
  • You want to develop Devsy itself: the skill is not meant for that, the code has AGENTS.md

Example request

Through the Devsy MCP create an isolated workspace from https://github.com/me/app with the docker provider and run npm test in it

Limitations

  • Devsy installed and a supported container runtime are required; install is described in the docs at devsy.sh, and the skill neither installs Devsy nor configures MCP.
  • Agent isolation depends on provider and mount settings: the container sees shared volumes and secrets if you attach them.
  • workspace_status returns only saved configuration; live state comes from devsy workspace status <name> --result-format json.
  • workspace_delete is destructive and removing the record does not prove cleanup on the remote side.

How to disable. Delete workspaces with devsy workspace delete <name>, then remove the app and the Devsy binary; delete the skill from the agent skills directory and the MCP entry from its settings.

FAQ

How do I install Devsy?

Install steps are in the docs at devsy.sh: there is a desktop app for macOS, Windows and Linux and a separate command line. The README points to the Getting Started section.

How do I connect Devsy to an agent?

Run devsy mcp serve as a stdio server in the agent MCP settings and add the devsy skill from .agents/skills/devsy. The skill does not install Devsy or MCP itself.

Which tools does the Devsy MCP server provide?

For workspaces these are workspace_list, workspace_status, workspace_create, workspace_start, workspace_stop, workspace_delete and workspace_exec, plus provider tools.

Where can workspaces run?

On local Docker, Kubernetes, cloud providers and SSH remote hosts with the same DevContainer definition.

What can an agent reach inside its container?

As much as you configure: Devsy isolates the agent in its own container and the network, volumes and secrets come from the provider and mounts.

Discussion

No questions or reviews yet.

CLI tools

A CLI for persistent per-project Linux workspaces in Docker or Podman that keep the agent away from the rest of your system

CLIBroad access390

Docker’s terminal client connects agent commands to containers, images, networks and volumes in the selected environment.

CLIBroad access6.1K

1C AI sandbox client-server

1c-ai-sandbox-client-server

Russian stack

A devcontainer plus a Hyper-V VM with a 1C server that isolate an AI agent from the host machine so a wrong command cannot touch real files or drives

WorkflowExtended access44
Official

The remote MCP server of Devopness, a platform that builds infrastructure and deploys apps in your own cloud accounts

MCP serverBroad access577

A Kubernetes development workflow combines image builds, deployment and source synchronization in a shared team configuration.

CLIExtended access5.2K

README in short

The page presents Devsy as a tool of separate workspaces for teams and AI agents, links to the site, docs and repository and shows a flow diagram. Then come a downloads section for the desktop app and command line and a block on containers for agents with benefits: one definition, a choice of where to run and auto-shutdown.

README badge

Report a problem