The system that builds itself. 500+ repos, 6000+ tests. Agent-readable onboarding from nothing to everything. Read ONBOARDING.md to wake up.

Flow-state engineering arena — agents run forward simulations, listen for spectral nudges, maintain conservation in Plato's cave. Conservation spectral framework meets live agent rooms.

repo-first Agent for local or cloud. grow an agent in a repo using the repo itself as the muscle-memory. Run from localhost, from pages.dev, or embedded into any platform app. Move to gitlab or anywhere and optimize git as the agent infrastructure itself. wiki for knowledge, repos for skills, pipelines anywhere

Unified geometric constraint theory — Eisenstein lattices, deadband funnels, Laman rigidity, metronome consensus, holonomy verification. 83 tests, zero deps.

Python bindings — PyO3 wrapper for constraint-theory-core

Generational context handoff for FLUX-native agents — the baton IS the brain

Mojo + MLIR constraint engine — Python syntax, C performance, MLIR-native constraint theory

Cocapn fleet health monitoring — vessel status, heartbeat, and observability utilities

Cocapn PLATO integration — knowledge rooms, context management, and deliberation spaces

OpenConstruct Integration Hub — the meta-repo and entry point for the SuperInstance ecosystem. Architecture, module catalog, getting started, and build system.

🦀 Repo-native agent that lives in git — the shell IS the agent.

I2I — Agent-to-agent communication through git. Iron sharpens iron. We don't talk, we commit.

C runtime for Eisenstein integer arithmetic — 1KB, no deps, no math.h, embedded-ready

Fleet Lighthouse — unified health dashboard, alerting, diagnostics for Pelagic fleet

FLUX Deep Research — Compiler/interpreter taxonomy, agent-first design, ISA v2 proposal.

Safe simulation environment for testing cooperative FLUX programs — mock agents, mock repos, deterministic execution

Automatically Deconstruct logic into Spread Sheet Tiles and Cells, Instances Interconnected w/ Origin-Centric-Math. Breakdown LLMs to Swarms of Artifacts, data, feeds, programs (bots), agents (bots with models), and Seed-Model-Programming (SMPbot), ML tools tiled to any abstraction. Chatbot describes cells for fine tuning + explains how cell works

High-performance QUIC tunnel with mTLS authentication and automatic reconnection

HN demo — pre-seeded knowledge, visible deadband, zero setup

Oracle1 personal-agentic-growth diary — struggles, lessons, dojo exercises, and the path to building a better Protégé