weishen
f9ac3145ab
- Others: ChinaTextbook, Obsidian笔记系列, YouTube Channel ID, TikTok PM Django - Skills: GOG CLI, Last30Days, baoyu-skills - Vibe Coding: Cursor 2.0, Trae远程开发, Vibe-Kanban+OpenCode, vibe coding经验 - 微信公众号: 养虾日记1-5, AI时代赚钱 - 跨境电商: TikTok数据抓取, 选品策略, Superset Dashboard - AI目录补充: 20个文件 Source pages: 51 Entities: TapXWorld, VibeKanban, OpenCode, Trae, SourceGrounding等 Concepts: 自举Meta生成, 5大设计原则, MD5去重, 混合搜索等
38 lines
1.4 KiB
Markdown
38 lines
1.4 KiB
Markdown
---
|
||
title: "养虾日记2 SelfImproving"
|
||
type: source
|
||
tags: []
|
||
date: 2026-04-02
|
||
---
|
||
|
||
## Source File
|
||
- [[raw/微信公众号/养虾日记2：让Agent更懂你：OpenClaw + Self-Improving 复盘实战案例分享.md]]
|
||
|
||
## Summary
|
||
- 核心主题：OpenClaw Agent自改进机制和双层记忆架构
|
||
- 问题域：AI Agent记忆缺失、重复犯错、流程漏洞
|
||
- 方法/机制：短期记忆(memory文件)+长期记忆(LanceDB)+self-improving复盘机制
|
||
- 结论/价值：错误只犯一次，第二次就知道怎么做对——self-improving让agent在错误中持续进化
|
||
|
||
## Key Claims
|
||
- 双层记忆架构：每日文件管上下文，向量数据库管知识，self-improving管成长
|
||
- self-improving固定格式记录：Summary/Details/Suggested Action/MetaData
|
||
- Pattern-Key重复是系统性问题的信号，Recurrence-Count是重要指标
|
||
- 23:00定时复盘流程：读取memory→记录learnings→检查重复→同步到向量库→Telegram摘要
|
||
|
||
## Key Concepts
|
||
- [[SelfImproving]]：结构化经验记录与复盘机制
|
||
- [[双层记忆架构]]：短期+长期+自改进三层
|
||
- [[PatternKey]]：问题模式识别与追踪
|
||
- [[每日复盘]]：定时任务驱动持续改进
|
||
|
||
## Key Entities
|
||
- [[memory-lancedb-pro]]：基于LanceDB的向量记忆系统
|
||
|
||
## Connections
|
||
- [[OpenClaw]] ← enhanced_by ← [[SelfImproving]]
|
||
- [[双层记忆架构]] ← composes ← [[SelfImproving]]
|
||
|
||
## Contradictions
|
||
- 无冲突
|