Current stageR12-R20 local foundations
Live development scope

What is in motion now

updated from the public state
  • R4.1-R10Protocol + conformanceGuarded / conforming
  • R12Worker boundaries + permission receiptsImplemented locally
  • R13A2A / ACP transport + adaptersImplemented locally
  • R14Model Atlas + usage ledgerImplemented locally
  • R15Control rooms + private projectionsImplemented locally
  • R16Library intake + progressive onboardingImplemented locally
  • R18Trusted devices + recoveryFoundation / gated
  • R19Authenticator + mobile approvalsFoundation / gated
  • R20Multi-authority queues + routingFoundation / gated
  • M5Hosted tenancy + external operationsNot enabled

Green means the local contract or conformance work is implemented. Yellow marks a foundation whose production boundary is still gated. Red means the capability is intentionally not enabled.

Control rooms

One governed record, different ways to work.

Interfaces may change. Authority does not. Every control room is a rebuildable projection of the same Relay record.

Integration philosophy

Local-first is the free reference path. Notion is a rebuildable knowledge projection. Slack is conversational coordination. GitHub is reviewable public-safe status.

R4 Shipped

Local-first

CLI, local archive, Library front end, Archive intake, progressive onboarding, Overseer round-table chat, and optional local models: the cheapest and most testable reference path when hosted tools are unavailable.

Allowed
Discover and ingest transcripts; catalogue the local Library; draft comparison packets; request review
Forbidden
Silent acceptance; model summary as decision; inventing authority outside validated records
Canonical?
No — local reference implementation; validated Relay records remain authoritative
R5 Planned

Notion

A structured knowledge and task projection for people who already work in Notion. Convenient and rebuildable—never canonical authority.

Allowed
Project tasks and knowledge views; notify; surface decision packets for a named human
Forbidden
Silent acceptance; checkbox or page text as approval; last-write-wins over Relay
Canonical?
No — rebuildable projection only
R6 Planned

Slack

Threads, notifications, review requests, and attributed commands. Messages can hand off work; they never equal approval by themselves.

Allowed
Notify; request review; thread coordination; explicit attributed command hand-offs
Forbidden
Message ≠ decision; emoji/reaction as acceptance; model-authored “approved” phrases
Canonical?
No — conversational projection only
R1–R4 Shipped contract

GitHub

Reviewable proposals, issues, receipts, and public-safe operational status committed to the durable Git record.

Allowed
Reviewable proposals; issues; receipts; public-safe status projection (live writes still gated)
Forbidden
Silent acceptance via label or bot comment; Issues UI as sole authority without Relay decision
Canonical?
Git (validated public commits) is the durable record; GitHub UI is not
R12–R16 Local foundation

AI tooling and Library

Tool schemas, connector readiness, conversation intake, Model Atlas evidence, and the AI Tooling & Integration Lab are local projections over bounded records.

Allowed
Read-only discovery; archive preview; schema and runtime evidence; public-safe release notes
Forbidden
Credential extraction; silent connector authorization; public transcript upload; model-authored decisions
Canonical?
No — the local Relay record and reviewed human decision remain authoritative
Boundary rules

Projections cannot create authority.

Allowed

Read, notify, draft, and project

Control rooms may surface tasks, request review, and present decision packets for a named human.

Forbidden

Silent acceptance by a model or channel

A Slack reply, Notion checkbox, or model summary cannot canonically accept consequential work without a human decision record.

Where it applies

Same record, different evidence thresholds.

Multi-model work

Compare competing outputs without losing who said what, what they used, or where they differ.

Operations

Bound incidents and changes with reproducible checks, visible exceptions, and accountable decisions.

Research

Preserve methods, artifacts, independent reproduction, adversarial review, and named scientific authority.

Governance

Keep evidence gaps, residual risk, review trails, and final approvals legible to the people responsible.