Token导航 LogoToken导航TokenDH.com
研究检索external-serviceclawhub未标认证来源可访问clear审计通过

auto-memory-keeper自动记忆保持器

Agent Skill

auto-memory-keeper 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

13,195

周安装

561

GitHub Stars

1

下载量

4,623
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:auto-memory-keeper(自动记忆保持器)
来源仓库:https://github.com/raydez/auto-memory-keeper
安装命令:
openclaw skills install auto-memory-keeper
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 auto-memory-keeper

简介

auto-memory-keeper 用于查找、检索和筛选相关信息,适合在 OpenClaw 中根据关键词快速定位候选结果时使用。

  • 支持每小时会议亮点捕获与每日记忆更新,适用于长期记忆维护。
  • 通过 clawhub 安装,命令为 openclaw skills install auto-memory-keeper,需结合来源仓库进一步确认具体用法。
  • 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操作。
  • 当前功能描述基于原始 README,实际能力以官方文档为准。

SKILL.md

name
auto-memory-keeper
description
Auto-capture hourly session highlights and update daily memory files. Keeps long-term memory continuous without manual recording. Triggered by cron or manual "remember this".
version
1.0.0

Auto Memory Keeper

Overview

Automatically extract session content and update daily memory files to maintain long-term memory continuity.

Two modes:

  • Passive mode: Cron job runs hourly (automatic)
  • Active mode: Manual trigger by user ("remember this", "log today")

How It Works

Triggers

  1. Auto-trigger: Cron job runs every hour
  2. Manual trigger: User says "remember this", "log today", "update memory"

Workflow

Step 1: Check Environment

DATE=$(date +%Y-%m-%d)
MEMORY_FILE=~/.openclaw/workspace/memory/${DATE}.md

if [ ! -f "$MEMORY_FILE" ]; then
  echo "# ${DATE} Log\
\
## Today's Items\
\
- " > "$MEMORY_FILE"
fi

Step 2: Fetch Recent Sessions

# Use sessions_list to get active sessions in last 60 minutes
# Use sessions_history to get session details

Step 3: Noise Filtering

Skip these message types:

  • Simple greetings ("hi", "hello", "hey")
  • Acknowledgments ("thanks", "ok", "sure", "yes")
  • Single word responses
  • Questions without context
  • HEARTBEAT_OK messages
  • System messages

Step 4: Extract Key Info

From filtered messages, extract:

TypeKeywordsFormat
Decision"decided", "chose", "adopted"- {time} {decision}
Tasknew feature, new config, new task- {time} {task description}
Progress"completed", "finished", "installed"- {time} completed {task}
Issueerror, bug, failed- {time} issue: {description}
Conclusionsummary, conclusion, finding- {time} {conclusion}

Step 5: Smart Deduplication

Check for duplicates before recording:

# Read current file content
# If similar content exists (>80% similarity), skip
# Otherwise append

Step 6: Update Memory File

Format:

## Today's Items

- {time} {item description}
- {time} {item description}
...

Examples

Input (user message):

"installed openclaw-tavily-search, score 3.639"

Stored:

- 21:47 installed openclaw-tavily-search skill (score 3.639)

Input (user message):

"auto-extract session content to memory file every hour"

Stored:

- 21:50 created auto-memory-keeper skill for hourly auto memory extraction

Configuration

Configurable via config.json:

{
  "memory_dir": "~/.openclaw/workspace/memory",
  "interval_hours": 1,
  "active_minutes": 60,
  "categories": ["Decision", "Task", "Progress", "Issue", "Conclusion"],
  "skip_patterns": ["^hi", "^hello", "^thanks", "^ok"],
  "dedup_threshold": 0.8
}

Quality Rules

  1. Concise - One sentence per item
  2. Timestamp - Always include time
  3. Dedupe - Avoid duplicate entries
  4. Categorize - Classify by type
  5. Skip noise - Don't record trivial messages

FAQ

Q: How to trigger manually? A: Say "remember this", "log today", or "update memory"

Q: Where is cron configured? A: Use cron job add command, see example in SKILL.md

Q: Where are memory files stored? A: ~/.openclaw/workspace/memory/YYYY-MM-DD.md

Testing

Periodically review memory files and adjust skip_patterns for better filtering.

Related

  • Works well with session-wrap-up (manual vs auto trigger)
  • Complex memories can be manually written to MEMORY.md (long-term)

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

76.49%
按下载量换算3,536

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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