
ephemora-cell
Capability-based WASM runtime for executing untrusted AI-generated code with enforced CPU, memory, time, I/O, and filesystem limits. Provides…

Capability-based WASM runtime for executing untrusted AI-generated code with enforced CPU, memory, time, I/O, and filesystem limits. Provides…

Proof-of-concept demonstrating a Node.js permission model bypass (CVE-2026-21636) that allows network access via undici/fetch to local services,…

Rootless container runtime and sandbox that launches kernel-enforced OCI images in milliseconds with no daemon, featuring resource profiles, seccomp…

Code for paper "ActBench: Self-Evolving Benchmark of Behavioral Safety in Cowork Agents"

Source code of a multiple series of tutorials about the hypervisor. Available at: https://rayanfam.com/tutorials

HyperDeceit is the ultimate all-in-one library that emulates Hyper-V for Windows, giving you the ability to intercept and manipulate operating system…

Macro-header for compile-time C obfuscation (tcc, win x86/x64)

PoC exploit for CVE-2026-64561, a KVM/x86 shadow MMU use-after-free enabling guest-to-host escape with kernel root code execution on the host.

Isolated JavaScript sandbox for Node.js that runs untrusted code with restricted access to built-in modules and host resources via Proxy-based…

Secure code execution

Sandboxed devcontainer for running Claude Code in bypass mode safely. Built for security audits and untrusted code review.

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

A secure low code deception runtime framework, leveraging AI for System Virtualization.

CVE-2020-0890 | Windows Hyper-V Denial of Service Vulnerability proof-of-concept code

Run untrusted AI code safely, fast

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


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