46 lines
2.3 KiB
Markdown
46 lines
2.3 KiB
Markdown
---
|
||
title: "Identity Graph Operator"
|
||
type: source
|
||
tags: [agent, the-agency, identity, graph, entity-resolution]
|
||
date: 2026-04-20
|
||
last_updated: 2026-04-20
|
||
---
|
||
|
||
## Source File
|
||
- [[raw/Agent/agency-agents/specialized/identity-graph-operator.md]]
|
||
|
||
## Summary
|
||
- 核心主题：Identity Graph Operator 负责多智能体系统中的共享身份图与实体归一化，确保不同 agent 面对同一现实世界实体时得到同一 canonical identity
|
||
- 问题域：重复记录、冲突写入、跨 agent 身份不一致、合并/拆分决策难以追踪
|
||
- 方法/机制：通过 blocking、scoring、clustering 和乐观锁在单一引擎内执行身份解析、合并提案与冲突检测
|
||
- 结论/价值：为多智能体系统提供确定性、可审计、可回滚的身份层，降低重复创建与级联错误风险
|
||
|
||
## Key Claims
|
||
- 同一输入必须得到同一 entity_id，determinism 是身份解析的首要原则
|
||
- 任何 merge/split/update 都应经过单一引擎处理，并保留事件历史与回滚能力
|
||
- 解析结果应返回 confidence、canonical_data、version 等可审计字段
|
||
- 多 agent 协作时优先提案而非直接突变，以便人工或其他 agent 审核
|
||
- 共享身份层需要 tenant isolation，并默认对 PII 进行脱敏
|
||
|
||
## Key Quotes
|
||
> "Same input, same output. Two agents resolving the same record must get the same entity_id." — 设计原则
|
||
|
||
## Key Concepts
|
||
- [[AI代理（Agent）]]：Identity Graph Operator 是支撑多智能体协作可靠性的基础设施
|
||
- [[Audit Trail]]：所有合并、拆分、匹配都应有理由和置信度记录
|
||
- [[Identity Governance]]：共享身份层的治理框架，强调 canonical identity 与权限边界
|
||
- [[Multi-Agent-System-Reliability]]：身份层是多智能体系统可靠性的前提之一
|
||
|
||
## Key Entities
|
||
- [[The Agency]]：开源 AI 智能体集合项目，Identity Graph Operator 所属项目语境
|
||
- [[Identity Graph Operator]]：该源文档对应的专门智能体角色
|
||
|
||
## Connections
|
||
- [[The Agency]] ← contains ← [[Identity Graph Operator]]
|
||
- [[Identity Graph Operator]] ← supports ← [[AI代理（Agent）]]
|
||
- [[Identity Graph Operator]] ← constrains_by ← [[Audit Trail]]
|
||
- [[Identity Graph Operator]] ← relies_on ← [[Identity Governance]]
|
||
|
||
## Contradictions
|
||
- 未发现明显冲突
|