Lab Notes
Living technical documentation for a local multi-agent systems lab — orchestration, MCP tooling, A2A agent-to-agent communication, multi-agent systems, reliability, evaluation and secure execution.
What the lab is
The lab is a local-first personal environment for experimenting with agentic AI workflows. It supports software development, research automation, media management, and personal knowledge organization through a set of specialized agents that operate under defined boundaries.
This documentation is the engineering reference for those systems. It is organized in ten sections that grow in depth, from a one-page summary to a per-component reference.
This documentation grows with the lab — expect new sections, deeper pages, and rewrites as the system evolves.
How to start
If you are new to the lab, follow the path that matches your role:
- Reviewer — read System Overview and Agents and Roles, then Reference Architecture.
- Operator — start with System Overview, then read Operations, Runbook, and Health Checks.
- Developer — start with System Overview, then read Tooling Layer, Tool Structure, and Project Templates.
- Auditor — start with Security Principles, then read Audit Model and Architecture Decisions.
The ten sections
10 · Architecture
System overview, components, data flow, framework.
20 · Agents
The five agents, their roles, their tools.
30 · Research
The research framework, phases, artifacts, providers.
40 · Tooling
CLI tools, MCP adapters, HTTP services.
50 · Templates
Project templates for new code.
60 · Operations
Operations, runbook, telemetry, health checks.
70 · Security
Security principles, audits, ADRs.
80 · Roadmap
Status legend, near/medium/long term.
90 · Appendix
Glossary, env vars, data dictionary, failures, ref.
What this site covers
The Coordinator Gateway and the Scout Gateway are two independent processes; requests that cross from one to the other — such as the Coordinator dispatching to the Research Worker or the Media Agent — are relayed over the A2A (Agent2Agent) protocol.