Skip to main content

Tier 2: Transformation

Open Repo Download ZIP

git clone https://github.com/AEEF-AI/aeef-transform.git

Goal: 4-6 week rollout with CI/CD, agent SDLC, and metrics.

The Transformation tier extends the Quick Start foundation with everything an organization needs to adopt AEEF standards across multiple teams over a 4-6 week period. It operationalizes the Transformation Track by providing runnable CI/CD pipelines, agent SDLC configuration, metrics collection infrastructure, and phased adoption tooling.

If you need an execution-first checklist instead of a capability overview, use Apply Path (Tier 2 Transformation).

What's Added Beyond Tier 1​

Everything from Tier 1: Quick Start is included, plus:

CapabilityStandards CoveredKey Files
Mutation testingPRD-STD-003Stryker / mutmut / go-mutesting configs
Full CI pipeline (7 stages)PRD-STD-003, 004, 007.github/workflows/ci.yml (expanded)
Documentation enforcementPRD-STD-005ADR templates, API doc generation configs
Technical debt trackingPRD-STD-006Complexity thresholds, debt labels in CI
Quality gatesPRD-STD-007Required status checks, branch protection
License compliancePRD-STD-008License allow/deny lists, SCA in CI
Agent SDLCPRD-STD-009Agent registry, contracts, handoff templates
KPI collectionFramework-wideProvenance generator, KPI schemas, collection scripts

Agent SDLC Overview​

The Transformation tier introduces the Agent SDLC Orchestration model with a starter configuration for multi-agent workflows:

  • Agent registry -- JSON manifest declaring each agent's identity, capabilities, trust level, and environment permissions.
  • Agent contracts -- Formal interface definitions specifying what each agent can read, write, and invoke.
  • Handoff templates -- Structured protocols for passing context between agents during SDLC phases.
  • Role prompts -- System prompts for each agent role aligned to the 11-agent model.
  • Skills catalog intake + mapping -- Quarantined external skill intake, attribution manifest, and AEEF-governed starter catalog mapping for controlled skill adoption.

See Agent SDLC Setup for the full configuration walkthrough.

Metrics Pipeline Overview​

The Transformation tier adds automated collection of AEEF KPIs as defined in the KPI Framework:

  • Provenance records -- Every CI run generates a provenance record capturing what was built, tested, and scanned.
  • KPI snapshots -- Weekly scripts aggregate CI data into KPI measurements (AI contribution ratio, defect density, security scan pass rate).
  • Schema validation -- All metrics records are validated against JSON schemas in CI to prevent data quality drift.

See Metrics Pipeline Setup for configuration details.

Operating Model Alignment​

The Transformation tier maps directly to the Operating Model Lifecycle:

Operating Model PhaseTier 2 Implementation
Business IntentStakeholder alignment templates, ROI calculator
AI ExplorationSandboxed experimentation configs, prompt library
Human HardeningCode review checklists, validation scripts
Governance GateCI quality gates, agent contract validation
Controlled DeploymentStaged rollout configs, canary analysis
Post-Implementation ReviewKPI dashboards, retrospective templates

Rollout Timeline​

Phase 1: Foundation (Weeks 1-4)​

  • Deploy Quick Start configs to 2-3 pilot teams
  • Set up CI pipelines with all 7 stages (lint, typecheck, test, mutation, SAST, SCA, license)
  • Install agent SDLC registry and initial contracts
  • Establish baseline KPI measurements
  • Train pilot teams on AEEF standards and AI tool usage

Phase 2: Structured Expansion (Weeks 2-3)​

  • Extend to all engineering teams
  • Enable mutation testing and enforce coverage thresholds
  • Deploy metrics pipeline and begin weekly KPI collection
  • Establish governance reviews for agent contract changes
  • Integrate documentation enforcement into CI

Phase 3: Enterprise Scale (Weeks 4-6)​

  • Enforce quality gates as required status checks across all repositories
  • Expand agent SDLC to multi-agent handoff workflows
  • Publish KPI dashboards for engineering leadership
  • Conduct maturity assessments targeting Level 3-4
  • Begin planning for Production tier if regulatory requirements apply

Repository​

github.com/AEEF-AI/aeef-transform

Language-Specific Guides​

  • TypeScript -- Full pipeline with ESLint, Stryker, and schema validation
  • Python -- Full pipeline with Ruff, mutmut, and pytest-cov thresholds
  • Go -- Full pipeline with golangci-lint, go-mutesting, and govulncheck
Last updated on by BalrogEg