
StepJack
Benchmarking framework for evaluating computer-use AI agents against multi-step indirect prompt injection, with automatic adversarial goal…

Benchmarking framework for evaluating computer-use AI agents against multi-step indirect prompt injection, with automatic adversarial goal…

Secure runtime to sandbox AI agent tasks. Run untrusted code in isolated WebAssembly environments.

Manages the core lifecycle of Qubes OS domains via a Python admin API, handling secure compartmentalization with Xen and exposing an event system for…

Docker-based sandbox for coding agents with isolated environments, preinstalled agent tooling, service control, and workspace bootstrap for secure…

Sandboxed Execution Environment

WASM sandbox with capability enforcement for AI agent code. Agents can only call explicitly provided tools with defined constraints. Sandboxed…

Lightweight, secure Linux sandboxes for untrusted processes. Runs in the browser and on the server.

Passive diagnostic tool that checks if a Linux system is vulnerable to CVE-2026-31431 by testing AF_ALG socket reachability, providing mitigation…

Sub-millisecond VM sandboxes for AI agents via copy-on-write forking

A local sandbox for your AI agents

Run Coding Agents in Sandboxes. Control Them Over HTTP. Supports Claude Code, Codex, OpenCode, and Amp.

Secure code execution

Let your AI go full send. Your home directory stays home.

Easily create full virtual machines that are sandboxed for development or computer use models.

A fuzzer for full VM kernel/driver targets

Ephemeral microVM sandbox for AI agents with network allowlisting, secret injection via MITM proxy, and VM-level isolation. Boots in under a second,…

Open-source automated malware analysis sandbox that runs suspicious files and URLs in isolated VMs and generates detailed behavioral reports.

Kernel-level eBPF sandbox for securing LLM agent tool calls made through the Model Context Protocol (MCP)