MANIFEST / MESH_NATIVE_APP

Distributed Credit Rating (DCR) — App Manifest

The complete Mesh-Native App manifest defining Distributed Credit Rating: a distributed financial intelligence application with encrypted local data, anonymized mesh scoring, and auditability. DOSFI is the constitution — no central authority ever sees raw financial data.

MANIFEST / HEADERApp Declaration

App ID

dosfi.app.distributed-credit-rating

Every actor and artifact is identity-rooted.

DCR binds the app, user, nodes, scores, and consent grants to attested @dosfi.ai identities. No anonymous scoring — every score is attributable to a signed consent bundle.

App Identity

dosfi.app.distributed-credit-rating

The DCR app is bound to an attested @dosfi.ai identity, signed by the DOSFI Identity Fabric at registration.

User Identity

<user>@dosfi.ai

Every user requesting a credit rating is bound to a hardware-attested @dosfi.ai identity. No anonymous scoring.

Node Identity

<node>@dosfi.ai

Nodes executing mesh scoring shards are bound to attested node identities with T2-verified or higher reputation.

Score Identity

UUIDv7 score_id

Every DCR score output receives a UUIDv7 identifier bound into the transport bundle envelope — fully attributable and auditable.

Consent Identity

signed consent_bundle

Every scoring request is bound to a cryptographically signed consent bundle recording the user's explicit authorization and selected privacy tier.

02 / CONSENT_MODELConsent Model

No scoring without explicit, revocable, tiered consent.

DCR enforces a strict consent model. No scoring occurs without a signed consent bundle. Consent is tier-specific, revocable, purpose-limited, and audit-logged.

03 / PRIVACY_TIERSPrivacy Tiers

Local-only by default. No central authority sees raw data.

DCR enforces a three-tier privacy model. Raw financial data is local_only by constitutional mandate — it never leaves the user's device. Only anonymized, normalized, aggregated vectors are mesh_ok. Cloud is a fallback for anonymized vectors only — never raw data.

local_only

Local-Only (Default)

Raw financial data, identity, account balances, payment history, and income history. Never leave the user's device. Encrypted at rest. Used to compute local features only.

Raw financial data
Identity
Account balances
Payment history
Income history

mesh_ok

Mesh-OK (Anonymized)

Anonymized, normalized, aggregated vectors safe for mesh distribution. Contains no raw financial data. Routed to attested nodes for scoring.

Anonymized vectors
Normalized features
Risk indicators
Trend metrics

cloud_ok

Cloud-OK (Fallback)

Fallback only — used when mesh is saturated, nodes timeout, or user chooses 'fastest' cost preference. Never used for raw data; only anonymized vectors.

Fallback only
Anonymized vectors only
Never raw data

04 / INPUTSScoring Inputs

Local-only raw data. Mesh-ok anonymized aggregates.

DCR uses local-only raw data to compute local features, and mesh-ok anonymized aggregates for mesh scoring. Raw financial data never leaves the device; only normalized, anonymized vectors are distributed.

Local-Only Inputs (never leave the device)

Encrypted on the user's device. Used to compute local features, not shared raw.

Income history
Payment history (rent, utilities, loans)
Account balances
Spending categories
Missed/late payments
Debt obligations
Savings behavior
Employment stability
Identity verification signals

Mesh-OK Inputs (anonymized + aggregated)

Safe to distribute — contain no raw financial data.

Risk features (normalized)
Behavioral vectors
Stability metrics
Trend indicators
Volatility scores
Debt-to-income ratio (anonymized)
Payment consistency score
Savings consistency score

05 / SCORING_PIPELINEThree-Stage Mesh-Native Scoring

Local extraction → mesh scoring → cloud fallback.

DCR uses a three-stage mesh-native scoring pipeline. Stage 1 runs entirely on the user's device. Stage 2 distributes the anonymized vector to the mesh. Stage 3 is a cloud fallback for anonymized vectors only — never raw data.

Stage 1 — Local Feature Extraction

local_only

Runs entirely on the user's device. Normalizes income, computes payment consistency, savings consistency, debt ratios, volatility, and stability, then builds a financial behavior vector. Produces a local-only feature vector — never shared raw.

Stage 2 — Mesh Scoring

mesh_ok

The anonymized vector is sent to the mesh. Multiple nodes compute partial scores; MeshInfer.AI aggregates results; Vital settlement pays nodes for compute. Privacy tier ensures no raw data is exposed. Produces the DCR score.

Stage 3 — Cloud Fallback

cloud_ok

Used only if the mesh is saturated, nodes timeout, or the user chooses 'fastest' cost preference. Cloud computes the score using the same model — only anonymized vectors, never raw data.

06 / SCORE_OUTPUTScore Output

Transparent, explainable, auditable.

Every DCR score output includes the score, risk tier, confidence level, a mesh-native explanation, telemetry, and a signed Vital settlement record.

Score

0–1000

Risk Tier

1–5

Confidence Level

0–1

Explanation

Mesh-native

Telemetry

Route, latency, savings

Vital Settlement Record

Signed

Example Score Output

DCR Score

742

Risk Tier

Tier 3 — Moderate

Confidence

0.87

Top Positive Factors

• 98% on-time payment rate
• Strong savings consistency
• Stable employment history

Top Negative Factors

• High debt-to-income ratio
• Elevated spending volatility in the last 90 days

Behavioral Insights

• Your financial stability trend is improving month-over-month.
• Reducing discretionary spending volatility could raise your score.

route: mesh_oklatency: 184mssavings: 92%

07 / RISK_TIERSRisk Tiers

Explainable, not opaque.

DCR maps scores to five transparent risk tiers. Each tier is defined by explainable factors — payment history, stability, volatility, savings behavior, and debt burden.

Excellent

800–1000Extremely low risk

Strong payment historyHigh stabilityLow volatilityStrong savings behavior

Good

700–799Low risk

Consistent paymentsModerate savingsStable employment

Moderate

600–699Medium risk

Occasional late paymentsHigher volatilityModerate debt

Fair

500–599Elevated risk

Irregular paymentsHigh volatilityHigh debt-to-income

High Risk

0–499High risk

Frequent missed paymentsUnstable incomeHigh volatilityHigh debt burden

08 / GOVERNANCEGovernance Rules

DOSFI is the constitution.

DCR is governed under the DOSFI Constitution. No central authority ever sees raw financial data. Scoring is explainable, auditable, and fairness-checked. Governance violations incur Vital penalties.

09 / VITAL_INTEGRATIONVital Integration

Compute settled in Vital (DIU).

Every mesh scoring shard and cloud fallback compute is metered and settled in Vital — the Distributed Intelligence Unit. Settlement is deterministic, signed by the Economics Adapter, and recorded on the Observability Fabric immutable audit ledger. Vital is a compute settlement unit, not a speculative instrument.

Mesh Scoring Settlement

Every mesh scoring shard is metered and settled in Vital (DIU). The Economics Adapter signs each settlement against the executing node's attested identity.

Node Participation Rewards

Nodes contributing compute to DCR scoring earn Vital proportional to verified TFLOPS, dispatch weight, and reputation tier.

Cloud Fallback Settlement

Cloud fallback compute is settled in Vital at a distinct rate, recorded with route=cloud_ok telemetry for full transparency.

Governance Penalties

Policy violations trigger deterministic Vital deductions recorded on the audit ledger — no discretionary fines.

No Speculative Tokens

Vital is a Distributed Intelligence Unit used strictly for compute settlement and governance accountability — not a speculative instrument.

10 / UNIFYING_THEORYUnifying Theory Compliance

Compliant with all 11 core components.

DCR is verified compliant with the Mesh-Native Unifying Theory of Distributed Intelligence — all 11 core components satisfied.

Distributed Inference Fabric

Compliant

Mesh scoring is sharded across the MeshInfer.AI runtime — multiple nodes compute partial scores, aggregated centrally by no one.

Mesh-Native Execution Model

Compliant

DCR executes on the DOSFI app runtime above the ABI boundary — zero install, no server deployment.

Workflow + State Machine

Compliant

Scoring requests are modeled as workflow bundles with three-stage state-machine transitions.

Device-as-Identity

Compliant

Users, nodes, and scores are bound to @dosfi.ai identities — no anonymous actors.

Local-First Privacy

Compliant

Default local_only privacy tier; raw financial data never leaves the user's device.

Deterministic Routing

Compliant

Anonymized vectors routed via Routing Adapter with trust, geo-compliance, capacity, and latency scoring.

Zero-Install Computing

Compliant

DCR streams UI schemas to any device — no installation required.

Cross-Device Intelligence

Compliant

Node readiness scoring evaluates local device capacity before scoring workload assignment.

Mesh-Native App Runtime

Compliant

Built above the ABI boundary — open surface, sealed core.

Mesh-Native Transport Layer

Compliant

All scoring artifacts transported as typed bundles through the five-adapter, seven-phase protocol.

Vital Settlement

Compliant

All mesh and cloud scoring compute settled in Vital (DIU) via the Economics Adapter — deterministic, audited, non-speculative.

11 / ELIGIBILITYRouting, Participation & Settlement Eligibility

Eligible for the full mesh lifecycle.

DCR is registered with the MeshInfer.AI scheduler and is eligible for mesh-native routing, open node participation, and Vital-backed compute settlement.

Eligible

MeshInference Routing

DCR is registered with the MeshInfer.AI scheduler. Anonymized scoring vectors are routed to attested nodes via deterministic dispatch-weight scoring.

Eligible

Node Participation

Any T2-verified or higher node may opt in to execute DCR scoring workloads. Nodes earn Vital for verified compute contributions.

Eligible

Vital-Backed Compute Settlement

Every executed scoring shard is settled in Vital through the Economics Adapter, with immutable audit records on the Observability Fabric.

REFERENCE / DEEP_DIVEExplore Foundational Components

Unifying Theory Identity Fabric Transport Layer Compliance & Security