agent.os/control-plane

Agent OS Atlas

From “I use agents like a chat” to “I run operational actors”

A practical, personal atlas for building a Personal Agent OS on Hermes Agent. Local-first, security always-on: the agent model, identity, a few long-lived profiles that delegate, native Kanban orchestration, skills, tools and MCP, memory, scheduling, and messaging only at the end.

30Notes
13Branches
12Phases
1Registries
The operating rule
Skill=procedure
Tool=capability
Profile=actor
Cron=time
Memory=continuity
Messaging=interface
Build route

Read it as a control plane, not a course catalog.

Start with the agent model, learn the Hermes runtime, set identity, then build the core: a few long-lived profiles that delegate, native Kanban orchestration, skills, tools, and memory. Add scheduling and messaging last. Security and evaluation run through all of it.

Open Start Here
00
Phase 00 · Get oriented

Orientation

Where to start, what to ignore for now, and how this atlas is organized.

Phase page
01
Phase 01 · Frame the agent

Mental Model

What a useful personal agent actually is, the observe→decide→act→verify loop, and graded autonomy with a human in the loop.

Phase page
02
Phase 02 · Know the runtime

Hermes Core

Hermes as a runtime: the platform-agnostic AIAgent, entry points, the agent loop, prompt assembly, the tool registry, sessions, and terminal backends.

Phase page
03
Phase 03 · Set the identity

Identity & Context

SOUL.md vs AGENTS.md vs context files: durable identity, operating instructions, and what does NOT belong in either.

Phase page
04
Phase 04 · Cast the actors

Actors & Delegation

The core branch — profiles vs subagents, the profile-vs-skill+delegation criterion, and a few persistent actors that delegate the heavy lifting.

Phase page
05
Phase 05 · Route the work

Native Orchestration

Use and extend Hermes's native Kanban orchestrator: route work to profiles by description, with approval gates and output contracts.

Phase page
06
Phase 06 · Capture procedure

Skills & Procedural Memory

Skills as procedural memory, the open standard, and the learning loop — letting the agent write its own skills, then curating them.

Phase page
07
Phase 07 · Wire capabilities

Capabilities

Core tools, toolsets, MCP, and plugins — the trust model and the line between read and write capabilities.

Phase page
08
Phase 08 · Keep continuity

Memory & Knowledge

MEMORY.md, USER.md, session search, and Honcho — what to remember, what to prune, and shared memory across actors.

Phase page
09
Phase 09 · Run on time

Scheduling & Automation

Cron as first-class agent tasks: fresh AIAgent per job, idempotency, delivery targets, and approval for anything sensitive.

Phase page
10
Phase 10 · Open the channel

Messaging Interface

Telegram as a control plane, only after everything is validated locally — pairing, allowlists, and securing dangerous tools.

Phase page
11
Phase 11 · Ship outputs

Practical Outputs

Automations that increase capacity to find, evaluate, answer, and produce sellable outputs — all in draft and approval.

Phase page
Phase ★ · Stay safe & honest

Always Active

Threat model, real isolation, draft/approval, and how you measure whether an actor actually helps before you automate it.

Phase page
Reference layer

Registries and operating memory

Registries keep external references, docs, tools, and source material normalized behind short atomic notes.