kvcache-ai/AgentENV

This page introduces an external open-source repository. It is not an HDATF product.

What it does

A distributed platform that runs many agent environments on Firecracker across machines and loads OCI images on demand. Environments can snapshot, pause, resume and fork into separate sandboxes for parallel agent work.

How it helps ATF

A reference for Harness on isolating each agent task in its own sandbox and on forking a running environment to try several paths in parallel. It is built for large-scale use on Linux, so its ideas may matter more than the platform itself.

License

MIT Permissive. Commercial use and changes are allowed if the copyright notice is kept.

More in this category

• yc-software/qm
  An agent harness for teams that runs in Slack and on the web. Each person and room gets scoped memory, files, permissions, scheduled jobs and a sandbox, and Pi, OpenCode, Codex or Claude Code can drive the same core.
• andrewyng/openworker
  An open-source desktop AI coworker that produces finished work such as security reviews, documents, Slack replies and a triaged inbox. It works with the user's chosen model or local Ollama, and every agent action is governed and logged.
• cloudflare/cloudflare-os
  An open-source AI productivity environment on Cloudflare Workers. It combines an agent chat preloaded with knowledge of how the company operates, sandboxed building and sharing of small apps, and a guardrail framework called Gatekeepers for agents and apps.
• citrolabs/ego-lite
  A macOS browser shared by a person and AI agents such as Codex or Claude Code. Agents run browser tasks in their own Spaces with the user's existing logins, while the user's own tabs stay separate.
• ifixai-ai/iFixAi
  A tool for auditing AI agents, run through a CLI or by the agent itself as a plugin or skill. It runs inspections that mix adversarial testing with operational checks, and returns a letter grade with a scorecard.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.