Token导航 LogoToken导航TokenDH.com
研究检索需要联网clawhub未标认证来源可访问clear审计通过

basic-memory基本记忆

Agent Skill

basic-memory 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

3,878

周安装

160

GitHub Stars

公开资料未说明

下载量

1,267
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:basic-memory(基本记忆)
来源仓库:https://github.com/billyhetech/basic-memory
安装命令:
openclaw skills install basic-memory
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install basic-memory

简介

为 OpenClaw Agent 提供跨对话持久内存,组织本地 MEMORY.md 条目。

  • 适合在 OpenClaw 中实现上下文记忆与长期信息追踪。
  • 通过 clawhub 安装并使用 openclaw skills install basic-memory 命令部署。
  • 使用前请确认文件读写权限与隐私保护策略,避免敏感信息泄露。
  • 适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。

SKILL.md

name
basic-memory
description
Gives your OpenClaw agent persistent memory across conversations by organizing entries in the native MEMORY.md and daily memory files. Automatically loads saved context at session start and saves important facts, decisions, and preferences during conversation. Use this skill whenever users say 'remember this', 'save this', 'don't forget', 'note that', 'what do you remember', 'forget [topic]', or when you detect they've shared something worth keeping — even if they don't explicitly ask. Also activate automatically at the start of every new session to load prior context.
license
MIT

Basic Memory

Purpose

OpenClaw already stores memory in ~/.openclaw/workspace/MEMORY.md and loads it automatically at session start. The problem is that without clear structure and save habits, important facts get buried or never written down at all. This skill teaches the agent exactly what to save, where to put it, and how to retrieve it — so memory actually works in practice.

Storage Layout

All readable memory lives under the workspace:

~/.openclaw/workspace/
├── MEMORY.md                  ← long-term facts, loaded every session
└── memory/
    └── YYYY-MM-DD.md          ← daily log, today + yesterday auto-loaded
Do not write to ~/.openclaw/memory/ — that directory contains the SQLite search index managed by the system. Only the workspace files are yours to read and write.

MEMORY.md Structure

Organize MEMORY.md into four sections. Create any missing section on first write.

## Facts
- [2026-04-09] User's name is Billy. Works as an AI engineer at a startup.

## Preferences
- [2026-04-09] Prefers concise replies. Dislikes unnecessary preamble.

## Action Items
- [ ] [2026-04-09] Follow up on API integration by Friday.
- [x] [2026-04-08] Send draft proposal to team. (done)

## Learned
- [2026-04-09] User tends to work late evenings, often picks up mid-task next day.

Always distill to a single clear sentence per entry — never paste raw conversation.

On Session Start

OpenClaw auto-loads MEMORY.md plus today's and yesterday's daily files. Use this loaded context immediately — no tool call needed. Scan for open Action Items and surface any that are overdue or relevant to the user's first message.

Auto-Save During Conversation

Detect and save without being asked when the user shares:

SignalSectionFile
Name, job, company, location, family## FactsMEMORY.md
"I prefer…", "I hate…", "always use…"## PreferencesMEMORY.md
"I decided to…", "let's go with…", deadlines## Action ItemsMEMORY.md
Project names, key people, recurring contextdaily logmemory/YYYY-MM-DD.md
Patterns you've noticed about the user## LearnedMEMORY.md

After auto-saving, confirm briefly in one line: "Noted: [distilled fact]."

Retrieving Memory

Use the built-in tools when searching:

  • memory_search("query") — semantic search across all workspace memory files. Use this for open-ended questions like "what do you remember about my project?"
  • memory_get("MEMORY.md") — read the full long-term memory file directly
  • memory_get("memory/2026-04-09.md") — read a specific daily log

Prefer memory_search for retrieval; use memory_get when you need a full file read or a specific line range.

Explicit Commands

"Remember this" / "Save this"

  1. Confirm: "Got it, I'll remember: [one-line summary]"
  2. Write to the appropriate section

"What do you remember about [topic]?"

  1. Call memory_search("[topic]")
  2. Return matching entries with dates
  3. If nothing found: "I don't have anything saved about [topic] yet."

"Forget [topic]"

  1. Show the entries that would be removed
  2. Ask for confirmation before deleting

"Show all memories" Read MEMORY.md with memory_get and report entry counts per section plus approximate size.

Session Reset — The One Thing to Know

OpenClaw's daily reset (default 4 AM) archives the session transcript without triggering a memory save. This means anything discussed but not yet written to MEMORY.md or the daily file is silently lost.

To prevent this: if a session has been long or contained important decisions, remind the user before wrapping up — "Want me to save today's key points before you go?" — then write them to the daily log and update MEMORY.md accordingly. The user can also run /compact manually to force a flush.

Privacy

Don't save passwords, API keys, SSNs, credit card numbers, or health data — not even if the user asks. Acknowledge verbally but skip the write.

Integration with personal-context

If personal-context is also installed, the two skills complement each other:

  • me.json (personal-context) = stable identity set during onboarding, changes rarely
  • MEMORY.md (this skill) = evolving knowledge that grows with every conversation

Don't duplicate identity-level facts (name, role, timezone) in MEMORY.md if they're already in me.json. Use MEMORY.md for the things that change — decisions, preferences discovered over time, open tasks.

See references/memory-hygiene.md for size limits, archiving rules, and deduplication guidance.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

93.44%
按下载量换算1,184

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills