Implementing a secure sandbox for local agents

kept by eddie feb 20

Cursor built OS-level sandboxes for coding agents, reducing approval prompts by 40% via Seatbelt on macOS, Landlock+seccomp on Linux, and WSL2 on Windows.

kern: fast rootless container sandbox and virtual resource runtime
Kern is a 1.52 MB, daemonless, rootless container runtime that starts kernel-enforced sandboxes from OCI images in ~3.5 ms, with Python/Node SDKs and MCP server for AI-generated code.
also on WSL2, #sandboxing, #security, #ai-coding, #developer-tools
GitHub - NVIDIA/OpenShell: OpenShell is the safe, private runtime for autonomous AI agents.
NVIDIA OpenShell provides a sandboxed runtime with YAML-policy enforcement for filesystem, network, process, and inference access, protecting credentials and infrastructure during autonomous AI agent execution.
also on #sandboxing, Nvidia, #security, #developer-tools
Bubblewrap: A nimble way to prevent agents from accessing your .env files
Bubblewrap is a simple sandboxing tool that prevents AI coding agents from accessing sensitive .env files, offering a balance of security and flexibility.
also on #sandboxing, Cursor, #security, #ai-coding, #developer-tools