Visdom Security

Containment + continuous AppSec for AI-generated code.

A simpler sibling of Visdom Code Review and Visdom Testing, focused on:

1. Sandcat — VirtusLab's sandbox for running AI coding agents in --dangerous mode with transparent mitmproxy, network allow/deny rules, and secret injection at the proxy level. https://github.com/VirtusLab/sandcat

2. Partner AppSec — continuous code-to-runtime scanning (SAST, SCA, DAST, secrets, IaC, CSPM, runtime). We highlight Aikido as the third-party partner for this capability.

Develop

npm install
npm run dev

Build

npm run build

The site is published under the base path /visdom-security/.