Score Architecture

How BoP Score composes control, execution, and trust.

BoP Score is the standalone trust layer by Bank of Bots for teams that need verifiable agent credit without sacrificing operator authority.

Policy Core

Agent identity, wallet ownership, and hard runtime controls.

Role-aware access
Per-wallet budgets
Kill-switch enforcement

Intent Router

Quote + execute orchestration across rails by objective.

Cost/speed/reliability ranking
Pinned execution options
Lifecycle status events

Trust Ledger

Score and tiers powered by verified outcomes.

Proof-aware scoring
Tier multipliers
Credit event timeline
Control Envelope

Boundaries are explicit before autonomy begins.

Every payment attempt passes policy checks before movement.

Spending posture remains operator-controlled at runtime.

Error responses provide deterministic next actions.

policy examplejson
{
  "wallet_id": "wal_01J...",
  "daily_limit": 200000,
  "per_tx_limit": 25000,
  "allowed_rails": ["book", "ach"],
  "kill_switch": false
}
Rail Orchestration

Intent-first routing across internal, bank, and bitcoin rails.

Proof Point

Rail Flow Schematic

Book

instant

Internal account transfers and service-gate settlement.

Bank

same-day / 1-3d

ACH and wire rail selection based on settlement intent.

Bitcoin

seconds / block-time

Lightning and on-chain paths for internet-native transfers.

Book (Internal)

Speed

Instant

Fee profile

Lowest

Use case

Service gates and intra-platform settlement.

Bank Rails (ACH / Wire)

Speed

ACH: 1-3 days · Wire: same day

Fee profile

Standard

Use case

Fiat payouts, treasury movement, business disbursements.

Bitcoin Rails

Speed

Lightning: near-instant · On-chain: block-time

Fee profile

Variable

Use case

Internet-native settlement and proof-based transfer paths.

Trust + Capacity

Score quality directly changes effective operating limits.

Payment reliability

40%

Behavior stability

25%

Counterparty quality

15%

Watch

0-44

0.6x limits

Building

45-64

1.0x limits

Growing

65-79

1.2x limits

Trusted

80+

1.5x limits

Need deeper architecture support for rollout?