Agent infrastructure
WindowsAgentArena
microsoft/WindowsAgentArena
A scalable Windows benchmarking platform for evaluating multimodal desktop agents with reproducible OS environments.
Topics
Related projects
qwen-code29.7kAn open-source terminal coding agent from Qwen that reads code, edits files, and runs commands across CLI, desktop, web, and editors.
learn-harness-engineering23.5kA project-based course teaching harness engineering for reliable AI coding agents through lectures and builds.
OpenSandbox16.5kAn isolated sandbox runtime for AI agents to run code, files, browsers, and desktops locally via Docker or on Kubernetes.
hermes-desktop14.5kA native desktop companion for installing, configuring, and chatting with the Hermes Agent in one place.
boxlite2.6kA lightweight hardware-isolated micro-VM sandbox runtime that gives AI agents safe, persistent code execution anywhere.
openclaw-android1.8kAn installer, compatibility patches, and Android app for running OpenClaw natively on phones without Linux.