Skip to content

Project

Docker Sandboxes

Docker feature that runs AI coding agents in isolated local or cloud microVM sandboxes, controlled via the sbx CLI, limiting access to the developer's machine.

Known aliases

  • Docker Sandbox
  • docker-sbx
  • Sandboxes
  • sbx
  • sbx CLI

Current stories

buildOne report1 publisher

Docker Sandbox kit confines a DeepAgents agent to one local model port

DeepAgents runs in a four-file Docker Sandbox kit whose agent can reach only a local Model Runner on port 12434, with no cloud keys. The egress policy is careful work, and exact reproduction still rests on what PyPI serves when each sandbox is created.

Publishers:dev.to

Reality

Evidence45
Adoption
Not scored
Hype gap+15
Incentives
Not scored
Confidence40
buildOne report1 publisher

Docker's sbx move shifts a coding agent's sandbox from laptop to cloud microVM in one command

Docker Cloud Sandboxes let one command move a coding agent's microVM sandbox from laptop to cloud, for runs Docker pitches at up to 21 hours. Where an agent runs becomes a CLI choice, though commenters argue the harder limit is which outside services it can reach.

Publishers:infoq.com

Reality

Evidence40
Adoption
Not scored
Hype gap+25
Incentives60
Confidence45
productOne report1 publisher

Docker hands CNCF a spec that makes an agent's permission list part of the image

Docker released its Sandbox Kit Spec under Apache 2.0 and is taking it to CNCF, with nine named vendors already shipping Kits for their own tools. Enforcement still belongs to the runtime. In the post, that runtime is Docker Sandboxes.

Publishers:docker.com

Reality

Evidence42
Adoption20
Hype gap+32
Incentives84
Confidence58
securityOne report1 publisher

CVE-2026-77179 let sandboxed agent code write anywhere the macOS host account could

Docker shipped the fix in Sandboxes 0.42.0 on September 7 and published the advisory on September 15. Every macOS build from 0.28.0 through 0.41.x is affected, and the default mount is the current directory, read-write.

Reality

Evidence70
Adoption
Not scored
Hype gap0
Incentives65
Confidence63
productOne report1 publisher

Docker's SBX kit writes a receipt for every eval command it runs

The open-source kit runs an evaluation command inside a Docker Sandbox and records six fields about what happened, which settles which command produced a number without settling what was installed at the time.

Publishers:docker.com

Reality

Evidence58
Adoption10
Hype gap+18
Incentives72
Confidence56
buildOne report1 publisher

Docker writes its own hypervisor, and inherits every bug in it

Docker VMM replaces the per-platform virtualization under Docker Desktop with one in-house backend. It is beta in v4.86 on Windows and macOS; Linux waits for October's general availability.

Publishers:cloudnativenow.com

Reality

Evidence32
Adoption24
Hype gap+28
Incentives72
Confidence46
productOne report1 publisher

GitHub's agent runner now has a trapdoor: Docker Sandboxes as a supported runtime

Agentic Workflows will start a coding agent inside a microVM with its own Docker daemon, so root-level freedom stops meaning root on the CI runner.

Publishers:docker.com

Reality

Evidence42
Adoption20
Hype gap+24
Incentives86
Confidence45
productOne report1 publisher

The allowlist read the command name, not what the command would do

CVE-2026-22708 let injected text rewrite a Cursor agent's environment, so an approved "git branch" ran something else. It worked with an empty allowlist too.

Publishers:docker.com

Reality

Evidence58
Adoption45
Hype gap+18
Incentives78
Confidence55
productOne report1 publisher

Pinned toolchain images were the easy part; scoping the agent is the new build problem

Docker's ESP32 walkthrough treats the espressif/idf image as settled ground and spends its energy on sandboxing AI coding agents. The containment boundary now matters more than the toolchain.

Publishers:docker.com

Reality

Evidence40
Adoption
Not scored
Hype gap+32
Incentives85
Confidence45