Semantic Index Contracts

LeafEnterprise publishes machine-readable semantic registries so future agents can understand system authority before they call tools.

Registry Files

File	Purpose
/machine/system-authority.json	Backend authority decisions, non-authoritative sources, and output boundaries.
/machine/data-plane.json	Source systems, storage layers, source roles, sensitivity, and consumer rules.
/machine/calculation-registry.json	Named savings/audit formulas, implementation paths, source rows, and blocked states.
/machine/evidence-lineage.json	Evidence-to-output chains for savings/audit, dashboard, workbook, PDF, AI, and agent tools.
/machine/backend-capability-map.json	Backend provisions, integrations, workers, SDKs, and presentation boundaries.
/machine/domo-integration.json	Domo PDP, group, App Studio/private embed, recovered estate evidence, and backend dual-enforcement rules.
/machine/api-reference.json	Live route families and endpoint metadata.

Agent Interpretation Rules

Agents should:

• read system-authority.json before selecting a tool,
• use data-plane.json to avoid treating the wrong source as authoritative,
• use calculation-registry.json before explaining or recomputing savings math,
• use evidence-lineage.json before making a guarantee, rebate, or dashboard-ready claim,
• use api-reference.json only after authority and evidence constraints are understood.

Validation

The semantic registries should remain small, explicit, and stable. They are not a data dump. They are an indexable contract layer over backend ownership, source boundaries, and reproducible output rules.