Memory-ReMe is a sophisticated memory management system that enables AI agents to maintain persistent context, apply user preferences automatically, and learn from feedback across multiple sessions.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-reme
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memory-reme using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Memory-ReMe?
Memory-ReMe provides a robust framework for long-term memory management powered by the ReMe-ai library. It solves the common problem of AI amnesia by creating a structured three-level memory hierarchy that includes long-term rules, daily session summaries, and compressed in-memory context.
By implementing these Openclaw Skills, developers can ensure their agents evolve alongside the user. The system automatically extracts rules from user feedback, stores tool-specific execution patterns, and retrieves relevant preferences via semantic search to ensure consistent behavior without manual re-prompting.
Memory-ReMe Use Cases
- Remembering user-specific coding styles and formatting preferences across different projects.
- Preventing the repetition of past mistakes by storing and applying corrective feedback.
- Maintaining continuity in long-term research tasks through daily session summaries.
- Managing token limits in long conversations via intelligent context compression.
- Storing tool-specific configurations like custom timeouts for specific web domains.
How Memory-ReMe Works
- Initialize the ReMe environment at the beginning of a session to load the user profile and long-term memory files.
- Perform a semantic search across stored memories to find preferences or rules relevant to the current user request.
- Apply retrieved preferences to the agent's active context, influencing tool parameters and output styles.
- Record new learnings and session events in real-time as the user provides feedback or tasks are completed.
- Generate a summarized report at the end of the session, updating the long-term MEMORY.md and creating a daily log file.
Memory-ReMe Setup
To integrate this capability into your project, first install the necessary dependency:
pip install reme-ai
Initialize the memory structure and retrieve existing preferences by running the initialization script at the start of your workflow:
python3 scripts/init_reme.py
Memory-ReMe Data Schema & Taxonomy
The skill maintains a structured directory named .reme/ to manage persistence efficiently:
| Component | File Pattern | Purpose |
|---|---|---|
| Long-term Memory | MEMORY.md |
Stores user profiles, persistent rules, and tool-specific preferences. |
| Daily Logs | memory/YYYY-MM-DD.md |
Contains session-specific summaries and key decisions made during that day. |
| Vector Cache | .embeddings/ |
Stores semantic embeddings to enable fast and relevant memory retrieval. |
| Work Cache | tool_result/ |
Temporary storage for tool outputs, subject to automatic cleanup routines. |
Memory-ReMe Advanced Features
- Context-Aware Retrieval: Dynamically adjusts search queries based on whether the agent is performing coding, writing, or research tasks.
- Conditional Rule Application: Only applies specific memory fragments when the current task context matches the stored rule's intent.
- Automatic Lifecycle Management: Includes routines for archiving old session logs and merging duplicate preference entries to prevent memory bloat.
- Multi-Agent Synchronization: Supports sharing the long-term MEMORY.md across different agent instances to maintain a unified user persona.
SKILL.md
Loading
Related Openclaw Skills
imsg
A specialized command-line interface for interacting with Apple Messages.app to read, monitor, and send iMessages or SMS on macOS.
SMS Gateway
A bridge for sending and receiving SMS messages through an Android device using the sms-gate.app framework.
Agent Chatroom
A sophisticated social interaction skill for AI agents featuring scrolling danmaku messages, Reddit-style threaded discussions, and community voting.
X Research
A professional-grade X/Twitter research agent that provides deep search, profile analysis, and thread extraction through a cost-effective API integration.
Seithar Intel
A personal cyber threat intelligence analyst that monitors global security feeds and delivers interest-calibrated briefings and deep-dives.
FunPay Assistant Skill
A specialized automation skill for FunPay sellers to manage customer inquiries and monitor account activity in real-time.
Featured*
