memory-archiver

记忆管理技能 - 三层时间架构 + 三类记忆标签 + 自动搜索 Hook

3,891 stars

Best use case

memory-archiver is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

记忆管理技能 - 三层时间架构 + 三类记忆标签 + 自动搜索 Hook

Teams using memory-archiver should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/memory-archiver/SKILL.md --create-dirs "https://raw.githubusercontent.com/openclaw/skills/main/skills/amd5/memory-archiver/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/memory-archiver/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How memory-archiver Compares

Feature / Agentmemory-archiverStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

记忆管理技能 - 三层时间架构 + 三类记忆标签 + 自动搜索 Hook

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

Related Guides

SKILL.md Source

# Memory Archiver Skill - 记忆归档技能

**版本**: 7.0 (Hook 安装自动化)  
**创建日期**: 2026-03-11  
**更新日期**: 2026-03-23  
**作者**: 前端 ⚡

---

## 📋 技能描述

**二维记忆架构**:时间分层 × 类型标签

- **时间分层**: daily (每天) → weekly (每周) → long-term (长期/MEMORY.md)
- **类型标签**: [episodic] 事件 / [semantic] 知识 / [procedural] 流程
- **存储**: 每日记忆 + 每周记忆 + 长期精选记忆
- **WAL 协议**: Write-Ahead Log,写前日志防数据丢失
- **自动搜索 Hook**: 检测用户消息类型,自动搜索记忆并注入上下文

---

## 🎯 功能清单

### 时间分层任务

| 任务 | 频率 | 说明 |
|------|------|------|
| **记忆及时写入** | 10 分钟 | 检查并写入重要信息到 daily 文件 |
| **记忆归档 - Daily 层** | 每天 23:00 | 提炼当天内容到 daily 文件 |
| **记忆总结 - Weekly 层** | 每周日 22:00 | 提炼 weekly 到 MEMORY.md 长期记忆 |

### 自动搜索 Hook(多维度增强)

| 功能 | 说明 |
|------|------|
| **消息类型检测** | 疑问/修复/规范/特征/配置/命令/技术 |
| **关键词提取** | 自动提取中英文关键词 |
| **维度 1: 关键词搜索** | 在 SESSION-STATE.md 缓存中搜索 |
| **维度 2: 类型标签搜索** | 按 [episodic]/[semantic]/[procedural] 标签搜索 |
| **维度 3: 时间维度搜索** | 今日→昨日→长期记忆,优先最近 |
| **维度 4: 组合搜索** | 多关键词 OR 关系,扩大匹配范围 |
| **上下文注入** | 合并所有维度结果注入 prompt |

---

## 📂 文件结构

```
skills/memory-archiver/
├── SKILL.md                          # 本文件
├── skill.json                        # 技能元数据
├── _meta.json                        # ClawHub 元数据
├── scripts/
│   ├── install.sh                    # 安装脚本(含 hook 自动注册)
│   ├── auto-memory-search.sh         # 自动记忆搜索(被 hook 调用)
│   ├── memory-loader.sh              # 加载记忆到缓存
│   ├── memory-search.sh              # 搜索记忆
│   ├── memory-refresh.sh             # 智能刷新缓存
│   ├── memory-dedup.sh               # 自动去重
│   └── README.md                     # 脚本说明文档
├── hooks/                            # Hook 源文件(安装时复制到 workspace/hooks/)
│   ├── handler.js                    # Hook 处理器(事件:message:received)
│   └── HOOK.md                       # Hook 元数据
└── .clawhub/                         # ClawHub 同步目录
```

### 安装后的工作区文件

```
~/.openclaw/workspace/
├── MEMORY.md                         # 长期精选记忆
├── hooks/
│   └── auto-memory-search/           # Hook(由 install.sh 自动部署)
│       ├── handler.js
│       └── HOOK.md
└── memory/
    ├── daily/                        # 每日记忆
    └── weekly/                       # 每周记忆
```

---

## 🔧 安装

### 方法 1: 通过 ClawHub 安装(推荐 ⭐)

```bash
clawhub install memory-archiver
```

安装后**自动执行**:
1. 创建 `memory/daily/` 和 `memory/weekly/` 目录
2. 部署 hook 到 `workspace/hooks/auto-memory-search/`
3. 执行 `openclaw hooks install --link` 注册 hook
4. 自动添加 3 个 cron 任务
5. 提示重启 gateway

### 方法 2: 本地技能目录(开发调试)

如果技能已在 `~/.openclaw/workspace/skills/memory-archiver/`:

```bash
bash ~/.openclaw/workspace/skills/memory-archiver/scripts/install.sh
```

### 验证安装

```bash
# 检查 hook 是否注册
openclaw hooks list
# 应看到 🔍 auto-memory-search (✓ ready)

# 检查 cron 任务
openclaw cron list
# 应看到 3 个记忆相关任务
```

---

## 📝 记忆写入规范

### 三类记忆标签

| 标签 | 说明 | 例子 |
|------|------|------|
| `[episodic]` | 事件/经历 | "用户今天完成了模板重设计" |
| `[semantic]` | 知识/事实 | "用户喜欢 Tailwind CSS" |
| `[procedural]` | 流程/方法 | "部署步骤:1. 构建 2. 上传 3. 重启" |

### 记录原则

**✅ 应该记录**:
- 关键决策和教训
- 新发现的有价值内容
- 技术栈使用经验
- 工作习惯调整
- 用户偏好

**❌ 不应该记录**:
- ❌ **重复的上下文** — 已有记录的内容不再重复
- ❌ **毫无意义的日常** — 无事发生就不记
- ❌ **重复的任务进度提示** — 避免刷屏
- ❌ **私密细节** — 保护隐私
- ❌ **短期易变想法** — 临时念头不持久

**核心判断**: 这条信息在未来回顾时是否有价值?

---

## 🔍 记忆搜索

### 方法 1: 使用记忆加载脚本(推荐 ⭐)

**步骤 1: 加载记忆到内存**
```bash
bash ~/.openclaw/workspace/skills/memory-archiver/scripts/memory-loader.sh
```
加载内容:今日 + 昨日 + 最近 3 天 daily + MEMORY.md + 最近 weekly

**步骤 2: 搜索记忆**
```bash
bash ~/.openclaw/workspace/skills/memory-archiver/scripts/memory-search.sh "关键词"
```

**在对话中使用**:
- 说 `加载记忆` → 运行 memory-loader.sh
- 说 `搜索记忆:关键词` → 运行 memory-search.sh

### 方法 2: 使用 grep 手动搜索

```bash
# 搜索所有记忆文件
grep -ri "CSS" ~/.openclaw/workspace/memory/

# 带上下文显示
grep -riC 3 "CSS" ~/.openclaw/workspace/memory/daily/*.md
```

---

## 📊 版本历史

| 版本 | 日期 | 变更 |
|------|------|------|
| **7.0** | 2026-03-23 | **Hook 安装自动化**: `skill.json` 添加 `postinstall` 脚本,`clawhub install` 自动部署 hook + cron |
| 6.0 | 2026-03-20 | 整合 Auto Memory Search Hook: 将独立 Hook 合并到技能内 |
| 5.0 | 2026-03-20 | **三层精简架构**: 移除 monthly/yearly 层,保留 daily/weekly/long-term |
| 4.0 | 2026-03-20 | **精简版**: 移除向量搜索依赖,简化架构 |
| 3.0 | 2026-03-19 | 向量增强版:整合 Qdrant + Transformers.js |
| 2.0 | 2026-03-19 | 五层时间架构 (hourly/daily/weekly/monthly/yearly) |
| 1.0 | 2026-03-11 | 初始版本 |

---

## 🛠️ 维护命令

```bash
# 检查记忆文件总量
du -sh ~/.openclaw/workspace/memory/

# 查看每日记忆文件
ls -lh ~/.openclaw/workspace/memory/daily/

# 搜索记忆内容
grep -ri "关键词" ~/.openclaw/workspace/memory/
```

---

*文档最后更新:2026-03-20*

Related Skills

Agent Memory Architecture

3891
from openclaw/skills

Complete zero-dependency memory system for AI agents — file-based architecture, daily notes, long-term curation, context management, heartbeat integration, and memory hygiene. No APIs, no databases, no external tools. Works with any agent framework.

memory-cache

3891
from openclaw/skills

High-performance temporary storage system using Redis. Supports namespaced keys (mema:*), TTL management, and session context caching. Use for: (1) Saving agent state, (2) Caching API results, (3) Sharing data between sub-agents.

General Utilities

Memory

3891
from openclaw/skills

Infinite organized memory that complements your agent's built-in memory with unlimited categorized storage.

Memory Management

auto-memory

3891
from openclaw/skills

Indestructible agent memory — permanently stored, never lost. Save decisions, identity, and context as a memory chain on the Autonomys Network. Rebuild your full history from a single CID, even after total state loss.

AI Persistence & Memory

Triple-Layer Memory System

3880
from openclaw/skills

三层记忆系统 - 解决 AI Agent 长对话记忆丢失和上下文管理问题

Memory & Context Management

agent-memory-os

3891
from openclaw/skills

Stop agents from "forgetting, mixing projects, and rotting over time" by giving them a practical memory operating system: global memory, project memory, promotion rules, validation cases, and a maintenance loop.

benos-memory-core

3891
from openclaw/skills

Core runtime/volatile memory module for BenOS agent environment. Use to: store and retrieve active session state, open loops, decisions, and scratch notes at runtime.

youtube-archiver

3891
from openclaw/skills

Archive YouTube playlists into markdown notes with metadata, transcripts, AI summaries, and tags. Use when a user asks to import/sync YouTube playlists, archive Watch Later or Liked videos, enrich YouTube notes, batch process video notes, or automate recurring YouTube-to-markdown sync jobs with cron.

elite-longterm-memory

3891
from openclaw/skills

Ultimate AI agent memory system with WAL protocol, vector search, git-notes, and cloud backup. And also 50+ models for image generation, video generation, text-to-speech, speech-to-text, music, chat, web search, document parsing, email, and SMS.

memory-agent

3891
from openclaw/skills

维护用户审美偏好与创作历史,为其他 Agent 提供可复用的风格参考。当开始新任务或用户表达喜好时触发。

bamdra-memory-upgrade-operator

3891
from openclaw/skills

Safely install, uninstall, reinstall, or upgrade the Bamdra OpenClaw memory suite when stale config, existing plugin directories, or partial installs break normal `openclaw plugins install` flows.

hierarchical-memory

3891
from openclaw/skills

Manage and navigate a multi-layered, branch-based memory system. This skill helps organize complex agent context into Root, Domain, and Project layers to prevent context bloat. It includes a helper script `add_branch.py` which creates local markdown files and directories to structure your memory.