PLATFORM LAYERS
Six layers. One trust plane.
Every autonomous action flows through six conceptual layers. Each layer produces artifacts consumed by the next — AEA → ATP → Forge → Axiom → AAECP → ACTA. The runtime services (Axiom, Verify, Trust, Core, and others) implement these layers.
How layers map to runtime services
Each conceptual layer is backed by one or more runtime services. The layer is what you reason about; the service is what runs.
AEA
What happened?
Ardyn Evidence Artifact — atomic, independently verifiable proof. Produced by Ardyn Core (issuance + verification) and the Verify Service.
Core: POST /v1/governance/check · Verify: POST /v1/verify/artifact
Kylewilson04/ardyn
ATP
Who should be trusted?
Ardyn Trust Profile — evidence-backed trust posture aggregated from AEAs over time. Implemented by the Trust Service.
GET /v1/agents/:id/trust-score
Kylewilson04/ardyn
Forge
Who or what may act?
Authority & Interlocks — principal identity, grants, delegation, quorum. Implemented by the Agent Registry and Tool Authority.
POST /v1/agents/register · POST /v1/tools/authorize
Kylewilson04/forge
Axiom
Under which policy?
Organizational Policy — evaluates rules against identity, trust state, and environment. ALLOW, BLOCK, ESCALATE, DEFER. Implemented by the Axiom Service.
POST /v1/governance/check · GET /v1/policies
Kylewilson04/ardyn
AAECP
How does work execute?
Governed Autonomous Execution — pre-execution authorization, post-execution verification, autonomous closure. Implemented by the AAECP Controller.
MCP intercept + governance check pipeline
Kylewilson04/forge
ACTA
How does the org learn?
Organizational Cognition — converts operational evidence into durable knowledge. Implemented by the Organizational Cognition service.
Audit exports, knowledge graph, cognitive queries
Kylewilson04/acta