The Valuation & Trust Overlay for Agentic Payments
PVM defines how AI agents, merchants, and payment facilitators express and consume value signals during agentic payments:
- 📊 Price Index: is this price fair?
- 🛡️ Credit Score: is this merchant trustworthy?
- ⭐ Feedback Aggregate: what do other buyers say?
- 🚦 Risk Level: should this payment be allowed?
PVM sits ON TOP of x402 / AP2 / Alipay / WeChat.
It does NOT replace payment protocols — it enriches them with machine-readable trust signals.
x402 V2 defines
maxAmountRequiredin the 402 response — but doesn't answer "is this amount reasonable?"
PVM fills this gap as a protocol-agnostic valuation overlay.
- Status: Active (v0.1.0)
- Scope: Core Value Metadata structure — embeds price index, credit score, and feedback into x402/AP2 payment responses.
- Files:
/schemas/v1/value-metadata.json— Core PVM schema/schemas/v1/price-index.json— Price index sub-schema/schemas/v1/credit-score.json— Credit score sub-schema
- Status: Draft / Community Feedback
- Scope: Defines a four-dimensional credit scoring model (300-850) for AI agents in payment scenarios.
- File:
/specs/credit-scoring-spec.md
- Status: Active
- Scope: Standardized metadata for AI service evaluation (pricing, performance, reputation, SLA).
- File:
/schemas/v1/ai-service-metadata.json
- Status: Active
- Scope: DDL for price index engine, merchant credit scoring, and user feedback tables.
- File:
/schemas/v1/price-index.sql
- x402 402 Response with PVM embedded 🆕 — Shows how x402 V2 payment response carries PVM trust signals
- Price Guard Hook — Fail-open price check before payment
- Schema Validator — Python validator for service metadata
- Weather API / Top-Up Service
- Service Metadata Schema (v1.0)
- Price Index + Credit Score JSON Schemas (v0.1.0) 🆕
- x402 V2 + PVM integration example 🆕
- Price Index SQL Schema (v1.0)
- Credit Scoring Standard Draft (v1.0-draft)
- Price Index Methodology Specification — Q3 2026
- Trust Protocol Interoperability Standard — Q4 2026
PVM is under active development. We are currently working on:
- Resolving weight inconsistency between credit-score.json and price-index.sql
- Defining risk_level assessment rules →
specs/risk-assessment.md - Publishing trust_score methodology specification →
specs/trust-score-methodology.md - Price Index Methodology Specification — Q3 2026
- Trust Protocol Interoperability Standard — Q4 2026
For detailed audit findings, see AUDIT-2026-08.md.
This is a community-driven effort. We welcome feedback, issues, and pull requests from developers, platforms, and researchers across the AI and payments ecosystem.
See CONTRIBUTING.md for guidelines.
- Read the draft specification
- Open an Issue for feedback
- Submit a Pull Request for improvements
- Join the Discussions
Maintainer: rhcjw
Together, let's build the trust layer for AI agent payments. 🛡️