Oz Dev Environments
https://github.com/warpdotdev/oz-dev-environments- Category
- AI Agents
- Rank
- No. 1955Tools index
Previous survey · No. 1966 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- warpdotdev
- GitHub
- 73 stars
- Date
About
Predefined dev environments for running Oz Cloud Agents and integrations from Warp.
What it does
A family of Debian-based container images supplies repeatable language toolchains for cloud agent jobs. You choose a focused stack or an all-in-one image, then attach repositories, setup commands, secrets, and runtime settings through Oz. Optional variants bundle several third-party coding agent command-line tools.
Why it's ranked here
This is a practical shortcut for teams already committed to Oz. The image matrix covers common language stacks, browsers, Docker client tooling, and two major Linux architectures. Its narrow scope is also clear: it standardizes containers, while Oz handles repositories, triggers, schedules, integrations, and orchestration.
What's good
Every image includes Node, Python, Git, curl, compiler tools, certificates, and Docker client utilities. Focused variants add Bun, Go, Rust, Java, .NET, Ruby, or browsers. Agent variants preinstall Claude Code, Codex, Gemini, and GitHub command-line tools. Published builds include software bills of materials and provenance.
Tradeoffs
The images contain Docker clients, Compose, and Buildx, but no Docker daemon. Third-party agent credentials must arrive through environment variables, preferably Oz secrets. Documentation and build configuration also disagree in places: the base image is described with Node 22, while its workflow inherits Node 24. The browser image installs Chromium behind a Google Chrome command alias.
How to use it well
Pick it when Oz jobs repeatedly need the same mainstream toolchain across command-line runs, integrations, schedules, and API calls. Reuse one environment per codebase, choose focused tags to limit image size, and reserve agent variants for alternate harnesses. It does not define the complete execution context or supply a Docker daemon.
Technical notes+
Dockerfile uses debian:trixie-slim, always installs Node, Python, uv, Git, build tools, certificates, and optional Docker client binaries, then gates language runtimes, browsers, and coding agents behind INSTALL_* build arguments. .github/workflows/build-images.yml builds a matrix of base, language, web, and full images, publishes regular and -agents tags to Docker Hub, targets linux/amd64 and linux/arm64 where specified, and enables SBOM plus maximum provenance. README.md documents Oz CLI, web app, API and SDK usage, integrations, schedules, secrets, custom images, and image extension. One notable mismatch is that README.md describes dev-base as Node 22, while the base workflow does not override the Dockerfile's Node 24 default.
Observed
- Packaging
- Prebuilt Docker images published under the warpdotdev namespace on Docker Hub
- Platform support
- Build matrix targets Linux amd64 and Linux arm64 for the documented image families
- Install surface
- Images are selected by Docker image tag when creating an Oz environment
- Interfaces
- Environments can be used through the Oz CLI, Oz web app, Agent API and SDK
- Base system
- Images derive from Debian trixie slim
- Build metadata
- Published workflow builds enable SBOM generation and maximum provenance
- Docker capability
- Docker CLI, Compose v2, standalone Compose, and Buildx are included without a daemon
Read from README.md, Dockerfile, .github/workflows/build-images.yml.
What it can do
Create predefined development environments
Configuration specifications → Ready-to-use development environment
Run Oz Cloud Agents
Agent code and dependencies → Executing cloud agent instances
Execute Warp integrations
Integration configurations and code → Running integration processes
Provision cloud-based development workspace
Environment requirements → Configured cloud development environment
Deploy agent-based applications
Agent application code → Deployed and running applications
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.