Composable agent runtime with enforced isolation boundaries
Build Software You Can Trust. Isolated, reproducible development environments and a secure package manager that give your whole team identical environments, while keeping AI agents off the laptop.
AI coding agents should work in attempts, not your working tree
Pi Agent 扩展，强制实现主/子 agent 完全隔离：独立进程、独立提示词、独立 skills、独立工具。主 agent 负责规划，子 agent 独立执行。
A workspace workflow tool for AI coding agents. Supports Git worktrees and jj workspaces, with fast isolated environment creation using ReFS block cloning on Windows.
MetaClaw fork: self-hosted GPU training (GRPO + LoRA + vLLM), per-agent skill isolation, multi-agent mode routing, training engine bug fixes. Train your AI agent on your own A100s. No cloud dependency.
Hands-on labs to reproduce real-world AI agent security exploits and compare isolation strategies: Docker containers vs micro-VMs
AI Agent Data Isolation System A universal framework to prevent data contamination in AI agent environments using vector databases. This system safeguards data integrity by detecting and filtering AI-generated content, applying session-based isolation, and preserving original data through configurable contamination detection patterns.
Add a description, image, and links to the agent-isolation topic page so that developers can more easily learn about it.
To associate your repository with the agent-isolation topic, visit your repo's landing page and select "manage topics."