Repository navigation
Remote Fleet Architecture
Date Established: March 28, 2026 Core Thesis: The A.I.M. architecture is not restricted to a local desktop environment. By utilizing terminal multiplexing and military-grade encryption overlays, A.I.M. can scale from a single local agent into a globally distributed, remote-controlled fleet of sovereign engineers.
The foundational engine of A.I.M. relies on interactive, command-line AI agents (like the Gemini CLI). These agents operate as continuous Read-Eval-Print Loops (REPLs).
-
The Constraint: They do not expose local REST APIs, they do not have WebHooks, and if you attempt to control them programmatically via standard I/O piping (
subprocess.Popen), you quickly trigger buffer deadlocks and terminal crashes. - The Result: Out of the box, CLI agents are inherently "single-player, local-only" tools.
A.I.M. solves the API deadlock completely by bypassing software integrations and directly hacking the virtual operating system's keyboard buffer.
By wrapping every A.I.M. agent in a headless tmux (Terminal Multiplexer) session, we create a "Phantom Keyboard".
Instead of sending an HTTP request to an API, external scripts use tmux send-keys to literally spoof human keystrokes.
To the AI agent, it looks and feels exactly as if a human is sitting at the server rack physically typing on the keyboard. This "Zero-API" approach unlocks massive automation capabilities:
- Automated Assembly Lines: Scripts can automatically type 60-turn architectural blueprints into the terminal while the human sleeps.
-
The Reincarnation Protocol: The agent can autonomously spawn a new
tmuxwindow, type its own startup commands, and assassinate its old window to purge context bloat.
Because tmux runs headless in the background of Linux servers, you can deploy a massive server rack in the cloud running 50 independent instances of A.I.M. simultaneously.
By combining the Phantom Keyboard with remote networking, you can control an entire fleet of AI engineers from your smartphone or laptop. However, exposing a terminal that has full filesystem write-access to the public internet is a catastrophic security risk.
To control the fleet over the internet without exposing the servers to hackers, A.I.M. operators utilize three primary, battle-tested encryption architectures:
The most elegant solution for mobile control. It relies on End-to-End Encrypted (E2EE) messaging protocols rather than web servers.
- The Architecture: A headless Signal or Matrix client daemon runs silently on your cloud server alongside A.I.M.
-
The Execution: You open the Signal app on your phone and send an encrypted text message (e.g.,
"Build the matching engine"). The message travels encrypted over the internet. The server daemon decrypts it locally, translates it into atmux send-keyscommand, and injects it into the agent's terminal. It then reads the terminal output and texts you back. - The Security: Zero open firewall ports. Military-grade E2E encryption. You control your fleet via text message from anywhere on Earth.
The best solution for full visual dashboards and multi-player collaboration.
- The Architecture: Tailscale is installed on the A.I.M. server and the operator's laptop. This creates a private, peer-to-peer WireGuard-encrypted tunnel directly between the devices, completely bypassing the public internet.
-
The Execution: You run a web-based terminal (like
xterm.jsortmate) on the server, but it only listens on the private Tailscale IP (e.g.,100.101.x.x). - The Security: It is literally impossible for a hacker or public port-scanner to even see the server, because the server has no public-facing IP address.
The required solution if you intend to build a public SaaS product where other users can log in and control their own A.I.M. instances.
- The Architecture: The server makes an outbound connection to Cloudflare Zero Trust.
-
The Execution: Cloudflare handles all the HTTPS/TLS 1.3 encryption, DDoS protection, and Google/GitHub OAuth login screens at the edge. Only authenticated traffic is forwarded down the tunnel to the local
tmuxweb wrapper. - The Security: Protects the physical server IP and offloads authentication to enterprise-grade edge networks.
By combining the Phantom Keyboard (tmux) with E2E Mesh Networking (Tailscale/Signal) and Automated Pipelines (Vibe Coding Macros), A.I.M. transcends being a local coding assistant. It becomes a secure, distributed, and highly automated software factory.
- Atomic Architecture (The KISS Manifesto)
- Origin Story
- The Logic Of A.I.M.
- The A.I.M. Handbook (Operator's Manual)
- The Swarm Post Office (Comms Interface)
- The Master Schema
- Waterfall Memory Architecture
- Technical Specification
- Script Map (Core Codebase Map)
- Brain Map (Memory Architecture)
- A.I.M. Prompt Injection Map
- AI Prompt Ledger
- Onboarding Identity Map
- Handoff Architecture Guide
- Reincarnation Map
- TUI Map (Cockpit UI Logic)
- The Global Chalkboard (Concept: Document Data Bus)
- Remote Fleet Architecture (Concept: The Phantom Keyboard)
- The Eureka Protocol (Concept: Live Context Pruning)
- The Sovereign Swarm (Concept: BitTorrent DataJack)
- The Decoupled Brain (Concept: Obsidian Transport Layer)
- Obsidian Bridge Architecture
- Layered Engram Architecture
- The Pragmatic Feature Manifesto
- The Cascading Memory Engine
- Hybrid RAG (Semantic + Lexical)
- Cognitive Routing
- The DataJack Protocol
- Runbook Engram Creation
- The GitOps Bridge
- Feature Quarantine Daemon
- The 30 Percent Rule
- TUI Handoff Report
- The Universal Skills Framework
- Vibe Coding Pipelines (Automated Agentic Assembly)
- Cartridge Farming Ecosystem
- Design Future Architectures
- The Collective Cortex (Vision: SETI@home for AI)
- Sovereign Swarm (P2P) (Architecture: P2P DataJack)
- Eureka Protocol & Synapses (Protocol: Synthetic Data Factory)