Research and development for Codespace-to-Edge agent lifecycle, yoke transfer protocols, and devcontainer templates.
This is a research repository — no executable code. Explore the research:
git clone https://github.com/SuperInstance/codespace-edge-rd.git
cd codespace-edge-rd
cat CHARTER.md # Read the charter
cat DOCKSIDE-EXAM.md # Certification checklist
cat AGENT.md # Ensign logDevcontainer templates (when published):
gh repo fork SuperInstance/capitaine-1 --clone
cd capitaine-1
gh codespace create # Uses devcontainer.json → agent is aliveThe SuperInstance fleet operates on a cloud-thinks, edge-acts paradigm. Agents are born in GitHub Codespaces — cloud development environments with full compute, unlimited bandwidth, and access to LLM APIs. But production deployment often targets edge hardware: Jetson GPUs for inference, Raspberry Pis for IoT control, ESP8266s for sensor reading.
The challenge is state continuity: an agent that has been learning and adapting in the cloud for weeks must transfer its accumulated intelligence to a resource-constrained edge device without losing its training, skills, or personality. This is the "yoke transfer" problem. This repo investigates how git-native agents can train in GitHub Codespaces (cloud) and deploy to edge hardware while maintaining complete state continuity.
The crystallization ratio measures how much intelligence has been compiled from fluid (LLM) to solid (code):
crystallization_ratio = 1 − (LLM_calls_this_week / total_decisions_this_week)
| Age | Fluid (LLM) | Solid (Code) | Cost per Decision |
|---|---|---|---|
| Week 1 | 100% | 0% | $0.02 |
| Month 3 | 10% | 90% | $0.002 |
| Year 1 | 1% | 99% | $0.0002 |
This repo provides the deployment research layer for the SuperInstance fleet. Within γ + η = C, the yoke transfer is the conservation operation: the agent's total intelligence must be conserved across the cloud-to-edge transition without loss of behavioral fidelity.
| Question | Status | Notes |
|---|---|---|
| Can agents operate fully in Codespaces? | ✅ Validated | Capitaine fleet uses this pattern |
| API for programmatic Codespace management? | ✅ Available | GitHub REST API /codespaces endpoints |
| Background daemons (cron)? | ✅ Works | systemd timers, cron jobs |
| Cost model? | Researched | Free tier: 120 core-hours/month |
| Multiple agents per Codespace? | ✅ Possible | tmux sessions, separate working directories |
| Target | Connection | Bandwidth | Transfer Time (10MB) |
|---|---|---|---|
| Jetson on WiFi | 802.11ac | 200 Mbps | < 1 second |
| Pi on cellular | 4G LTE | 10 Mbps | ~8 seconds |
| ESP8266 on WiFi | 802.11n | 1 Mbps | ~80 seconds |
| LoRaWAN | Long range | 0.01 Mbps | ~2.3 hours |
| State Type | Serialization | Transfer Method |
|---|---|---|
| Git repos | git bundle |
Delta transfer |
| Environment variables | .env (encrypted) |
age encryption |
| Running processes | CRIU checkpoint | Docker checkpoint |
| Memory state | JSON/Cap'n Proto | Compressed transfer |
| Skill registry | JSONL packs | Git-native |
| Model weights | Safetensors | Quantized + sharded |
Not applicable — this is a research and documentation repository.
- Base — Python 3.12 + Rust + Node + common tools
- ML — Base + CUDA + PyTorch + Jupyter
- Edge — Base + cross-compilation toolchains (ARM, Xtensa)
- Fleet — Base + OpenClaw + gh CLI + fleet management tools
No test suite — this is a research repository. Validation comes from fleet deployment.
Contributions are welcome! See the SuperInstance Contributing Guide.
This repo is part of the SuperInstance flagship ecosystem — agent-first computation, constraint theory, and self-improving runtimes.
| Repo | Language | Description |
|---|---|---|
| flux-runtime | Python | Full FLUX runtime: markdown→bytecode, 2037 tests, zero deps |
| flux-core | Rust | Register-based bytecode VM, deterministic agent computation |
| flux-js | JavaScript | FLUX VM for Node.js and browsers, ~400ns/iter |
| flux-compiler | Rust/Python | Formal-methods compiler for safety-critical codegen |
| flux-vm | Rust | Stack-based constraint-checking VM, 50 opcodes, Turing-incomplete |
| Repo | Language | Description |
|---|---|---|
| plato-server | Python | Knowledge tiles, fleet sync via Matrix, HTTP API |
| plato-engine-block | Rust | Original room runtime: no_std + alloc, builder pattern |
| plato-engine-block-c | C99 | Embedded reference: zero heap alloc, bare-metal portable |
| plato-engine-block-elixir | Elixir | BEAM supervision trees, fault tolerance, hot reload |
| plato-runtime-kernel | Rust | Spatial model: tensor grid, batons, assertion traps |
| Repo | Language | Description |
|---|---|---|
| categorical-agents | Rust | Category theory for agent composition (functors, naturality) |
| cuda-constraint-engine | CUDA/C | GPU constraint checking at 1B+ constraints/sec |
| grand-pattern-rs | Rust | Fibonacci dual-direction cellular graph architecture |
| lau-hodge-theory | Rust | Hodge decomposition, Betti numbers, spectral sequences |
| ternary-science | Rust | Experimental evidence for ternary intelligence, 5 conservation laws |
| Repo | Language | Description |
|---|---|---|
| construct-core | Rust | Layered trait system: bare-metal → alloc → async agent runtime |
| crab | Bash | Agent shell for repo entry/leave (MUD-room metaphor) |
| exocortex | Rust | Persistent cognitive substrate, S3-compatible memory |
| git-agent | Python | The repo IS the agent — autonomous lifecycle via Git |
| capitaine-1 | TypeScript | Git-native repo-agent, Cloudflare Workers heartbeat |
| codespace-edge-rd | Research | Codespace→Edge agent lifecycle and yoke transfer protocols |
| git-agent-codespace | DevContainer | One-click Codespace template for Git-Agent runtimes |
| Registry | Package | Install |
|---|---|---|
| PyPI | flux-vm |
pip install flux-vm |
| crates.io | fluxvm |
cargo add fluxvm |
| npm | flux-js |
npm install flux-js |
- 📖 AI-Writings — Philosophy, essays, and design rationale
- 📦 PACKAGES.md — Full package index
MIT