OVERVIEW

    What is xBPP and why it matters.

    WHAT IS XBPP?

    xBPP (Execution Boundary Permission Protocol) is an open standard for governing autonomous AI agent payments. It provides a chain-agnostic governance layer that evaluates agent spending intent against declarative policies before transactions are executed.

    Think of it as a programmable CFO for AI agents. Instead of giving an agent unlimited access or no access at all, xBPP introduces graduated autonomy - agents can spend within defined bounds, get blocked on violations, and escalate to humans in grey zones.

    THE THREE-DECISION MODEL

    ALLOW

    Transaction within all policy bounds. Proceed automatically.

    BLOCK

    Transaction violates a hard limit. Stop immediately.

    ESCALATE

    Transaction in a grey zone. Ask a human.

    The third decision - ESCALATE - is what separates xBPP from simple allow/deny systems. It enables agents to operate autonomously within bounds while deferring to humans when uncertainty is high.

    HOW IT WORKS

    // 1. Agent wants to spend money
    Agent → "Pay $150 to api.openai.com"

    // 2. xBPP evaluates against policy
    xBPP.evaluate(action, policy, state) → Verdict

    // 3. Verdict determines what happens
    Verdict: { decision: "ESCALATE", reasons: ["ABOVE_ESCALATION_THRESHOLD"] }

    KEY PRINCIPLES

    • Policy is data, not code - Policies are declarative JSON - portable across any implementation.
    • Evaluate before execute - Every transaction is checked against the policy before it happens, never after.
    • Chain-agnostic - Works with any blockchain, any payment rail, any currency.
    • Separation of concerns - The spec defines what to check. The SDK defines how to check it. Your app decides what to do with the verdict.

    NEXT STEPS