# AEGIS — Operational Maturity Model

- **Document ID:** ARCH-30
- **Phase:** D — Deployment & Long View
- **Status:** Draft for review (post four-reviewer discipline)
- **Version:** 0.1
- **Date:** 2026-07-10
- **Owner:** Chief Security Architect

---

## 1. Purpose

Define the operational-maturity levels AEGIS-as-a-product progresses through, per axis. Complements ARCH-27 (technical Assurance Levels) with an *operational* view: how the humans + processes + tooling around AEGIS evolve.

## 2. Maturity Axes and Levels

Five axes, five levels each. Levels are progressive — a level implies all of the prior.

### 2.1 On-Call

- **M-OC-1 Ad hoc.** Anyone reachable responds.
- **M-OC-2 Rotation.** Documented rotation; primary + secondary.
- **M-OC-3 Coverage.** Regional coverage or follow-the-sun; sustainability tracked.
- **M-OC-4 Specialization.** Domain-specific escalation (Kernel, AI, Recovery, Ops).
- **M-OC-5 Autonomous triage.** Automated first-response for common categories; humans focused on novel.

### 2.2 Drills

- **M-D-1** Table-tops only.
- **M-D-2** Runbook drills quarterly.
- **M-D-3** Chaos + runbook drills monthly.
- **M-D-4** Cross-team + tenant-facing drills biannually.
- **M-D-5** Red/purple team engagements + industry-standard exercises annually.

### 2.3 Ceremonies

- **M-C-1** Documented but not rehearsed.
- **M-C-2** Rehearsed annually.
- **M-C-3** Rehearsed biannually + drilled-with-failure-injection.
- **M-C-4** Cross-team rehearsal; multi-region participation.
- **M-C-5** Ceremony automation for safe parts + human-in-the-loop for authority.

### 2.4 Metrics & Feedback

- **M-M-1** Ad hoc dashboards.
- **M-M-2** Standard SLO tracking (ARCH-24 baseline).
- **M-M-3** Continuous validation + drift monitors + adaptive thresholds.
- **M-M-4** Trend-based decision-making at the executive level.
- **M-M-5** Peer benchmarking + attack-simulation coverage metrics.

### 2.5 Incident Response

- **M-IR-1** Reactive.
- **M-IR-2** Structured lifecycle (ARCH-21) with runbooks.
- **M-IR-3** IC rotation + blameless post-mortems + corrective-action tracking.
- **M-IR-4** Cross-team drills + jurisdictional matrix + tenant-communication SLO.
- **M-IR-5** Automated response for narrow categories + tenant-side integration + industry threat-intel sharing.

## 3. Target Trajectory

| Milestone | On-Call | Drills | Ceremonies | Metrics | IR |
|---|---|---|---|---|---|
| **v0-GA** | 2 | 2 | 2 | 2 | 2 |
| **v0.2** | 2 | 3 | 2 | 3 | 3 |
| **v1** | 3 | 3 | 3 | 3 | 3 |
| **v1.5** | 4 | 4 | 4 | 4 | 4 |
| **v2** | 4 | 5 | 4 | 5 | 5 |
| **v3** | 5 | 5 | 5 | 5 | 5 |

Deltas between actual and target are Debt Register entries (ARCH-29).

## 4. Progression Criteria

Advancement between levels requires:
- Evidence: sustained metric compliance over ≥ 90 days.
- Documentation: runbooks + rehearsal artifacts.
- Auditor sign-off.
- Debt Register updated.

## 5. Independent Architecture Review

- **F-1.** *Progression can stall on culture.* Auditor tracking + trend visibility surface stalls.
- **F-2.** *Levels can be gamed.* Evidence-based criteria + auditor examination reduce gaming.

## 6. Adversarial Architect Review

- **A-1.** *Insider inflates maturity claim.* Auditor + evidence + external audit annually.
- **A-2.** *Attacker times an attack for pre-level-progression phase.* Cross-team drills verify progression is real.

## 7. Operational Reliability Review

- **O-1.** Level progression sustainable — not permanently ratcheting.
- **O-2.** Regressions are OK — drop a level with documented cause; don't hide.

## 8. Self-Critique

- **S-1.** *Levels are opinionated.* Framework is starting point; per-tenant customization possible; industry-benchmark alignment planned.
- **S-2.** *"Coverage" and "specialization" (M-OC-3/4) depend on headcount.* Small teams cap at M-OC-2 — documented as sustainable.
- **S-3.** *Automation of ceremonies (M-C-5)* is risky.* Only safe steps automated; authority remains human. Detailed in Phase E.

## 9. First-Target Analysis and Redesign

**Target:** self-reported maturity levels. If teams over-report, downstream planning falsely assumes readiness. Response:
1. **Evidence-based progression** (§4).
2. **External audit** verifies annually.
3. **Metric anchors** — level claims tied to specific metric SLOs; regressions auto-decrement claim.
4. **Auditor visibility** on all level changes.

## 10. Decisions

### D-30-1. Five-axis, five-level maturity model with per-milestone targets

- **Reason.** Explicit operational trajectory over 15 years.

## 11. Change Log

- **0.1 (2026-07-10)** — Initial draft.