# Agent Receipts

> Cryptographically signed audit trails for AI agent actions

The Agent Receipts protocol is an open standard for creating tamper-evident records of AI agent actions. Each receipt is a W3C Verifiable Credential, signed with Ed25519, and hash-chained into a verifiable sequence.

## Key links

- Specification: https://agentreceipts.ai/specification/overview/
- Getting Started: https://agentreceipts.ai/
- GitHub: https://github.com/agent-receipts
- Full context for LLMs: https://agentreceipts.ai/llms-full.txt

## Repositories

- spec: Protocol specification and JSON Schema
- sdk-ts: TypeScript SDK for creating and verifying receipts
- sdk-py: Python SDK for creating and verifying receipts
- openclaw: OpenClaw plugin for automatic receipt generation

## Core concepts

- Agent Receipt: A signed record of a single agent action (W3C VC with type AgentReceipt)
- Receipt Chain: Hash-linked sequence of receipts forming a tamper-evident log
- Action Taxonomy: Hierarchical action types (filesystem, system — with more domains planned) with risk levels
- Risk Levels: low, medium, high, critical — implementations may escalate but not downgrade
