Contra Collective / Applied Systems Lab

Applied research in
intelligent systems.

The Lab studies practical questions in agent infrastructure, developer tooling, and web systems. Each project begins as a working hypothesis, then is tested, refined, and released when it proves useful.

lab_observations.log Active

$ lab observe --projects all

01 ContraCore in development
02 OpenAstra online
03 Ingenium online
04 RiftPanel online
05 ShellShade online
5 systems observed · hypotheses tested · one release in preparation_
[ 01 ] Selected Builds

Systems built for practical use.

Production-minded experiments spanning agent infrastructure, web frameworks, and developer workflows.

01 / AI-native intentional framework
Coming soon

ContraCore

An AI-native, intentional front-end framework: a schema-first system for building interfaces and their full-stack boundaries with fewer valid patterns. Routes, data, actions, forms, and access rules derive from explicit contracts; scaffold and agent-callable verify tools keep both humans and coding agents inside one canonical architecture.

Schema-firstAgent verificationOwn SSRNode · Edge · Static
quick_start01

$ release channel: preparing

◌ validation in progress

Public release in preparation
02 / Self-hosted agent runtime
Available

OpenAstra

Production infrastructure for agentic systems: durable memory, multi-channel routing, tool execution, and swarm coordination in one deployable runtime.

Node 20+Self-hostedOpen sourceAgent runtime
quick_start02

$ npx openastra init

✓ ready

03 / Modern HTTP framework
Available

Ingenium

An Express-shaped TypeScript framework with returned responses, lazy parsing, and an incremental migration path for applications that need to evolve without a rewrite.

Node.jsTypeScriptIncremental migration
quick_start03

$ npm create ingenium

✓ ready

04 / AI-native desktop workspace
Available

RiftPanel

A persistent command center for teams running multiple coding agents: durable sessions, stable layouts, and terminal output that stays navigable through long tasks.

DesktopDeveloper toolsAI workflows
quick_start04

$ agents: 4 · sessions: live

✓ ready

05 / Open-source utility
Available

ShellShade

A focused shell utility released from our internal toolkit. Small surface area, clear documentation, and designed to do one job without becoming a platform.

Open sourceShellGitHub Pages
quick_start05

$ source ./shellshade

✓ ready

[ 02 ] Operating Principles

How an internal fix becomes a real product.

01

Built from friction

Every project begins with a recurring constraint we encountered while shipping real systems.

02

Used before announced

We run the tools ourselves, find the sharp edges, and harden the workflow before publishing it.

03

Kept intentionally small

A narrow tool with a legible model is easier to trust, maintain, and adapt than another platform.

Have a stubborn systems problem?

The next lab project could start with yours.

Bring us the constraint that keeps surviving every workaround. We will help determine whether it needs an integration, a product, or a better architecture.

Start a conversation →