AI GOVERNANCE / CONTROL PLANE

v0.8.0 close-ready next: v0.8.1

AI governance for services that cannot afford ambiguity. Control the logic centrally. Keep execution local.

`v0.8.0` turns the governance surface into something you can actually read and compare. Risk becomes a signal. Service offers become explicit commitments. Specialist composition becomes a concrete routing choice.

Release

v0.8.0

final renewal

Solution Tiers

3

audit -> setup -> monitoring

Debt Classes

4

observable and reviewable

Specialists

5

recommendation only

RELEASE STATUS HERO

v0.8.0 / control-plane locked

SIGNAL NODE

Close Verdict

clean renewal and control-plane extension.

Boundary

  • central governance = policy, registry, recommendation
  • service runtime = deploy, domain, monitoring execution
  • no real MCP runtime hosted in the central kernel

Rendered Surface

  • static docs deploy
  • JSON-driven panels
  • Markdown-driven operating prompts

Motion Budget

  • single reactive signal canvas
  • viewport-gated reveals and pulses
  • reduced-motion path included

GOVERNANCE TRACE SIMULATOR

Choose the offer. Watch the operating burden change.

Pick the commercial tier and the debt posture. The page will show which execution obligation, review shape, and specialist MCP composition become reasonable.

Service Offer

Debt Posture

Selected Offer

tier-2 keeps onboarding temporary, but governance setup becomes part of the promise.

Execution Burden

The independent service still owns deployment and runtime, but now it must expose a governed kernel and structured plan draft.

Recommended Specialists

governance-api ui-ux frontend-perf

SOLUTION TIER MATRIX

Three offers. Three accountability levels.

Read these as commitment structures, not marketing packages. The further you move down the grid, the more explicit the receipts, ownership, and monitoring posture become.

tier-1 / audit only

Point-in-time review, risk register, and handoff recommendation.

tier-2 / governance setup

Audit plus governance bootstrap, structured plan draft, and gate posture.

tier-3 / monitoring link

Audit and governance setup plus persistent monitoring-link receipts and cadence.

COGNITIVE DEBT CONSOLE

Turn agent risk into enterprise-readable debt.

Hallucination is not a mystical property here. It becomes a named debt class with signals, mitigations, and a decision the buyer can actually live with or reject.

hallucination

Claims without replayable evidence stay as debt, not trust.

omission

Missing packets, receipts, or rollback context become explicit debt items.

SPECIALIST MCP ROUTING BOARD

Central governance recommends. Independent services compose.

The registry answers a practical question: which specialist should enter the loop right now, and why? Runtime implementation still lives outside this kernel.

governance-api

Receipt lookups, registry hydration, and status summaries.

ui-ux

Narrative, hierarchy, and interaction calibration.

MONITORING LINK RAIL

Tier-3 is where governance stops pretending it is just onboarding.

Persistent linking requires explicit owner route, evidence freshness, uptime expectations, and review cadence. The promise gets heavier because the relationship gets real.

uptime

error surface

evidence freshness

owner route

review cadence

monitoring_link_receipt review_cadence_receipt owner_route_receipt

VERSION EVOLUTION + PROMPT SURFACE

Docs evolve with the version. Prompts stay public-readable on purpose.

JSON keeps the checkpoints and mirrors. Markdown carries the role system, visual direction, and next-version pressure. The gap between contract and atmosphere is intentional.

Current close

v0.8.0 can close on its control-plane extension and docs renewal.

Next target

v0.8.1 becomes the next pressure point for specialist composition and debt posture.

landing-narrative

Public copy stays tied to release truth.

mission-control-experience

The site behaves like an operator console, not a pitch deck.