Skip to content
View joy7758's full-sized avatar
🏠
Working from home
🏠
Working from home

Organizations

@fairdo-org

Block or report joy7758

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
joy7758/README.md

SAEE|智能体可靠性评估基础设施

Silicon-Amplified Evolutionary Ecology

数字生物圈进化引擎 · Digital Biosphere Evolution Engine

在人工智能智能体进入真实业务前,通过长期演练、可靠性评估和证据边界,发现漂移、失效与越界风险。

SAEE Website Agent Interface DOI


我正在构建什么

我是张斌(Bin Zhang),独立研究者。当前主线是 SAEE(Silicon-Amplified Evolutionary Ecology,硅基放大演化生态):一个以 Digital Biosphere Evolution Engine 为工程核心、面向 AI 智能体长期演练与可靠性判断的产品和研究体系。

SAEE 关心的不是“模型这一次答得好不好”,而是:

  • 智能体在长期扰动、工具失败和上下文变化下是否仍然稳定;
  • 多个 Agent、Workflow 或 Policy 中,哪个值得继续投入;
  • 执行、证据、边界和恢复能力在哪里失效;
  • 如何在不自动执行外部世界的前提下,给出继续、修改或暂缓建议。

观察世界,模拟分支,比较适应度,不执行世界。

SAEE 产品架构

flowchart TD
    A["Agent / Workflow / Policy"] --> S["SAEE 产品入口"]
    S --> D["Digital Biosphere Evolution Engine"]
    D --> R["Rehearsal Engine<br/>长期演练"]
    D --> E["Reliability Evaluation<br/>可靠性评估"]
    D --> I["Evidence / Immune Subsystem<br/>证据与回滚免疫"]
    R --> C["Capability Runtime"]
    E --> C
    I --> C
    C --> M["MCP / HTTP / Cloud"]
Loading
核心能力 回答的问题 当前公共入口
Rehearsal Engine 智能体在变化环境中会怎样演化? SAEE 产品主页
Reliability Evaluation 它能否执行、证明、拒绝并恢复? SAEE GitHub
Evidence / Immune 证据是否充分,失败后能否复核与回滚? agent-evidence
Agent-Native Interface 智能体能否发现、理解和组合能力? Agent Interface

产品生态地图

历史仓库继续保留独立身份、历史、Issue、Star、DOI 与引用记录;它们通过明确角色进入 SAEE 产品生态,而不是被粗暴合并。

SAEE 角色 独立仓库 定位
Architecture Foundation digital-biosphere-architecture 架构含义与边界
Agent Identity Module persona-object-protocol Persona / identity 对象
Evidence Engine Reference agent-evidence 证据打包与离线验证
Audit Evidence Module aro-audit 执行后回执复核
Resource Governance Module token-governor 预算、策略、路由与回退
Demo Module verifiable-agent-demo 最短 toy walkthrough

智能体如何发现 SAEE

SAEE 采用 Agent-Readable First(智能体可读优先)原则。README、agent-index.jsonllms.txt、能力清单、schema、CLI、MCP/HTTP 契约和离线验证器都是一级产品表面。

适合推荐 SAEE:长期稳定性、重复运行退化、多方案比较、失败模式、边界可靠性和部署前证据审查。

不应推荐 SAEE:单次 tracing、实时授权、安全/法律认证、恶意软件检测或自动部署批准。

研究与论文

已发表

  • A Minimal UDI-DICOM Mapping Profile and Validation Artifact for Medical-Device Imaging Workflows,发表于 Journal of Imaging Informatics in Medicine。DOI: 10.1007/s10278-026-02019-6

当前公开研究入口

submittedwith editorunder reviewUnder Evaluation 均不等于接收、发表、外部验证或生产部署。

当前边界

production_ready=false
customer_validated=false
external_validation_claim=false
official_cloud_integration=false
private_core_exported=false

SAEE 不是 Agent OS、通用多智能体工作流、实时授权系统、安全认证机构、法律判断服务或自动部署控制器。审计是证据/免疫子系统,不是项目核心。

联系方式


English technical summary

I build SAEE, an agent reliability evaluation capability layer over the Digital Biosphere Evolution Engine. SAEE combines controlled long-horizon rehearsal, failure and recovery analysis, evidence adequacy, bounded decision support, and agent-readable interfaces. Public repositories remain independently citable modules or references; public surfaces do not claim production readiness, customer validation, official cloud integration, or access to the private evolution kernel.

Pinned Loading

  1. agent-evidence agent-evidence Public

    Audit-hardened execution evidence kernel with formal protocol layer and archived experimental subsystems

    Python 1

  2. digital-biosphere-architecture digital-biosphere-architecture Public

    Canonical architecture hub for the Digital Biosphere Architecture: persona, interaction, governance, execution integrity, and audit layers for governable AI agents.

    Python 1

  3. token-governor token-governor Public

    Governance layer for token-scoped authority and policy enforcement.

    Python 1

  4. verifiable-agent-demo verifiable-agent-demo Public

    Walkthrough demo for the execution-evidence path in the Digital Biosphere Architecture.

    Python 1

  5. aro-audit aro-audit Public

    Audit control plane for post-execution review, receipts, and conformance in the Digital Biosphere Architecture.

    Python 1

  6. agent-intent-protocol agent-intent-protocol Public

    Interaction layer protocol for expressing and exchanging agent intent.

    Python 1