Kubernetes-sigs Agent Sandbox

Docker Sandboxes is a Docker-developed proprietary tool for running AI coding agents and other workloads inside isolated microVMs on a local machine or Docker-managed cloud infrastructure. Each sandbox has its own kernel and filesystem, controls access to the host filesystem, network, and tools, provides a private Docker daemon for test containers, uses secret placeholders, blocks network access by default, and records an audit trail. It supports agents including Claude Code, Codex, Gemini CLI, and OpenCode, with configurable network and filesystem rules, read-only mounts, MCP catalogs, and policy-based governance controls. It is distributed as the `sbx` command-line tool for macOS, Windows, and Linux, with stable, release-candidate, and nightly builds.

Microsandbox is a local-first microVM runtime and library for running untrusted workloads, including AI-agent tasks, user code, plugins, CI jobs, development environments, scrapers, and automation. It uses hardware-level isolation and runs standard OCI container images from registries such as Docker Hub and GHCR, with Docker-like image, command, shell, and volume workflows. The project provides SDKs for TypeScript, Rust, Python, Go, and Ruby, along with a CLI for booting and controlling sandboxes. Applications can create microVMs as child processes without a setup server or long-running daemon; sandboxes can also run detached for long-lived sessions. Agent Skills and an MCP server support agent-created sandboxes, and the project documents secret handling intended to keep keys out of the VM. It runs on Linux, macOS, and Windows, with platform-specific virtualization requirements, and is described as beta software.

kagent is an open-source Kubernetes-native framework for running AI agents as cluster workloads, developed for DevOps and platform engineering teams. Agents are defined as Kubernetes custom resources and deployed with kubectl or GitOps; kagent provisions their runtime, connects them to tools through MCP, and can coordinate agents through A2A. It supports Go and Python runtimes, multiple LLM providers and agent frameworks, persistent vector-backed memory across sessions, context compaction for long conversations, human approval gates, sandboxing, RBAC, and OpenTelemetry traces, Prometheus metrics, and structured logs for prompts, tool calls, and results. The project is distributed under the Apache 2.0 license and is a CNCF Sandbox project.

Agent Substrate is an open-source runtime and control plane for running agent-like workloads at high density. It manages the lifecycle of sandboxed actors, including creation, destruction, suspension, and resumption, and assigns them to a smaller pool of worker pods while routing traffic to the active workers. The system uses Kubernetes for infrastructure provisioning and pod lifecycle management, with support for sandbox technologies including microVMs and gVisor. It preserves actors' volatile memory and filesystem state through full-state snapshots, enabling sub-second suspend and resume operations and multiplexing many stateful actors across shared infrastructure. It manages standard OCI containers at the kernel level and is designed to support different agent frameworks and harnesses, including ADK-compatible actors, LangChain agents, coding environments, and MCP servers. Agent Substrate is intended for running agents at scale rather than building them, and its workloads do not have to be literal AI agents. The repository notes that the project is not an officially supported Google product.