Local-first verification layer for AI agents: evidence, provenance, policy gates, human approval, and auditable Trust Receipts before a claim, memory write, or action is trusted.接入配置该项目没有提供统一的启动命令,请按源码仓库的说明安装。