Step 5 · Knowledge graph

Live instances of entities, authorities, decisions.

Steps 1–4 define the contracts. Step 5 is the state: a live, instance-level graph of every entity, every signed delegation, every capability grant, every scope, every state vector, every decision and every evidence pack the runtime has produced — typed under the KYE™ ontology and bound by signatures.

rocket_launch Apply for pilot → menu_book Read the docs

1 · What lives here

Authority Graph™ · Decision Map™ · Evidence Graph™ · Blast Radius Map™.

KYE-Graph™ ships four normative graph projections over the same instance store. Each is a typed view, not a separate database.

account_treeAuthority Graph™Entities, delegations, authorities, capabilities, scopes — every edge typed by the ontology.

mapDecision Map™The signed, replayable artefact emitted with every authorize call. Inputs → rules → outcome.

verifiedEvidence Graph™Audit events, evidence packs, transparency receipts — bound to the decisions that produced them.

device_hubBlast Radius Map™The downstream-derived authorities a single revocation or quarantine signal would reach. Pre-emit; helps operators choose the right action.

2 · Why a graph and not a table

Authority is a chain, not a row.

Tabular ACLs answer "can actor A do X?". The agentic stack needs "can actor A act on behalf of principal P, under delegation D, attenuated by scope S, in state vector V, against capability C, with evidence E?" — a path query, not a row query. A graph is the only honest data model. The Authority Graph™ is what the policy engine traverses in step 6.

Where to go next

Continue the stack →

Every claim here is backed by the open KYE Protocol™ contracts and verifiable end-to-end from the publisher's JWKS — you check it yourself, you don't take our word for it.

schema ← Step 4 · Schemas gavel Step 6 · Policy engine →

Ready to see your AI agents flagged?

Start in shadow mode. We’ll deliver your first Evidence Pack™ in 4–8 weeks.

rocket_launch Apply for pilot → science Try the sandbox →