← Back to Skills Marketplace
49
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install depth-memory
Description
深层记忆系统。处理知识积累、技能经验、长期档案类任务时触发。提供关键字检索、记忆存取、索引更新功能。
README (SKILL.md)
Deep Memory（深层记忆）
分层记忆体系的第三层（MEMORY.md → deep-memory → 互联网）。处理知识、技能、经验类任务时优先使用。
记忆层次
memory/ & MEMORY.md ← 日常/高频上下文（第一层）
deep-memory/ ← 知识/技能/经验档案（第二层）← 当前层
互联网搜索 ← 无匹配时 fallback
工作流
查询（遇到问题时）：
- 读取
DEEP-MEMORY.md（索引文件） - 从用户输入中提取关键字，匹配索引表的 Keywords 列
- 命中 → 读取对应 deep-memory 文件
- 无命中 → 继续互联网搜索
新增（整理有价值内容时）：
- 内容存入
deep-memory/YYYY-MM-DD HH_MM_SS.md - 提取关键字并更新
DEEP-MEMORY.md索引
目录结构
deep-memory/ ← 记忆文件目录
DEEP-MEMORY.md ← 索引（表格形式）
skills/deep-memory/
SKILL.md ← 本文件
scripts/
query.mjs ← 关键字检索脚本
add.mjs ← 新增记忆脚本
list-index.mjs ← 列出所有索引条目
索引格式
DEEP-MEMORY.md 为 markdown 表格，列：Keywords | Description | File Path
| Keywords | Description | File Path |
| :--- | :--- | :--- |
| @k3000/build | mjs打包工具 | `deep-memory/2026-04-27 13_06_20.md` |
脚本用法
# 查询关键字（返回匹配的文件路径）
node scripts/query.mjs \x3C关键字>
# 新增记忆（交互式）
node scripts/add.mjs
# 列出所有索引
node scripts/list-index.mjs
触发条件
遇到以下类型任务时，主动使用本 skill：
- "我的 xxx 库怎么用来着"
- "之前那个 xxx 配置在哪个文件"
- "整理一下今天学到的 xxx"
- "我有没有关于 xxx 的记忆"
- 任何知识/技能/经验类问题
Usage Guidance
Install this if you want a local long-term memory layer. Before using it, be aware that entries saved under deep-memory and indexed in DEEP-MEMORY.md may influence future answers, so avoid saving passwords, tokens, private personal data, or anything you do not want persisted.
Capability Analysis
Type: OpenClaw Skill
Name: depth-memory
Version: 0.1.0
The 'depth-memory' skill implements a local knowledge management system for the agent to store and retrieve long-term information. The provided Node.js scripts (add.mjs and query.mjs) perform basic file operations on markdown files within the workspace and do not exhibit any signs of data exfiltration, malicious execution, or unauthorized access.
Capability Assessment
Purpose & Capability
The skill's stated purpose matches its behavior: it reads a local DEEP-MEMORY.md index, retrieves matching memory files, and can add new local memory entries.
Instruction Scope
Instructions tell the agent to proactively use the memory layer for knowledge, skill, and experience questions; this is purpose-aligned but means stored memories may influence later answers.
Install Mechanism
There is no install spec, but the skill documents Node.js script usage and references a list-index script that is not included in the manifest.
Credentials
File access is scoped to local deep-memory files and the DEEP-MEMORY.md index, which is proportionate for a memory skill.
Persistence & Privilege
The skill intentionally creates and updates persistent local memory files; users should avoid storing secrets or sensitive personal data unless they want it reused later.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install depth-memory - After installation, invoke the skill by name or use
/depth-memory - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
Initial release of deep-memory skill:
- Introduces a structured, multi-layered memory system for managing long-term knowledge, skills, and experience.
- Supports keyword-based querying, memory storage, and index updating.
- Provides dedicated scripts for querying, adding, and listing memory records.
- Uses a clear directory and index structure for efficient retrieval.
- Designed to trigger automatically during knowledge, skill, or experience-based tasks.
Metadata
Frequently Asked Questions
What is Deep Memory?
深层记忆系统。处理知识积累、技能经验、长期档案类任务时触发。提供关键字检索、记忆存取、索引更新功能。 It is an AI Agent Skill for Claude Code / OpenClaw, with 49 downloads so far.
How do I install Deep Memory?
Run "/install depth-memory" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Deep Memory free?
Yes, Deep Memory is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Deep Memory support?
Deep Memory is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Deep Memory?
It is built and maintained by qshfu (@qshfu); the current version is v0.1.0.
More Skills