A sophisticated hierarchical memory system that enables AI agents to retain long-term context and personal growth without session bloat.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install rey-memory
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 rey-memory 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 Rey Memory?
Rey Memory provides a structured approach to long-term memory for AI agents, allowing them to transcend individual chat sessions. By implementing a multi-layered storage strategy, it ensures that your agent evolves based on supervisor feedback and past experiences. This system is designed to keep context windows lean by compressing older data while promoting core values and critical insights to permanent storage.
Developing persistent digital personalities requires more than just logging; it requires a system that understands what to remember and what to forget. By utilizing this skill within the ecosystem of Openclaw Skills, developers can ensure their agents maintain a consistent self-identity and build deeper relationships with their users over time through a logical progression of short-term to core memory layers.
Rey Memory Use Cases
- Maintaining a consistent personality and self-identity across multiple weeks of interaction.
- Recalling specific supervisor instructions or unique preferences from past sessions.
- Tracking an agent's technical learning curve and growth milestones.
- Managing long-term project context without exceeding LLM token limits.
- Building trust through the retention of shared emotional moments and high-level summaries.
How Rey Memory Works
- At the conclusion of each session, the agent analyzes the conversation to extract key learnings, insights, and identity-related updates.
- Summarized data (5-10 lines) is saved to the Daily Memory layer within a connected Google Sheet.
- Periodic compression routines run weekly and monthly to synthesize daily logs into high-level highlights, deleting expired raw data to prevent bloat.
- Truly critical insights or changes in belief systems are promoted to Core Memory for permanent retention.
- Upon starting a new session, the agent automatically retrieves core memories and the most recent 7 days of daily logs to reconstruct its state.
Rey Memory Setup
To deploy this memory architecture for your agent, follow these steps:
- Create a Google Spreadsheet named
{AGENT_NAME}-Memory. - Define the necessary sheets:
daily_memories,weekly_memories,monthly_memories,core_memories, andgrowth_log. - Ensure your agent has authorized access to the Google Sheets API via your environment configuration.
- Install the skill into your project using the CLI:
openclaw install rey-memory
Rey Memory Data Schema & Taxonomy
The memory system organizes data across several specialized sheets to optimize retrieval and compression. This structure is a standard for agents using Openclaw Skills.
| Sheet Name | Key Columns | Retention Policy |
|---|---|---|
| daily_memories | date, summary, learnings, emotions, raw_quotes | 30 Days |
| weekly_memories | week, highlights, growth, challenges | 1 Year |
| core_memories | id, content, category (value/belief), importance | Permanent |
| growth_log | date, before_state, after_state, trigger_event | Permanent |
| conversation_index | id, date, channel, topic, memory_layer | Permanent |
Rey Memory Advanced Features
- Importance-based scoring (1-10) to automate the promotion of data to permanent core memory.
- Supervisor-specific dialogue prioritizing, ensuring interactions with project leads are weighted more heavily.
- Automated context window management that caps memory injection at approximately 450 lines.
- Multi-platform indexing to track memory origin across Discord, CLI, and web interfaces.
- Self-identity synchronization that updates the agent's core system prompt based on validated growth logs.
SKILL.md
Loading
Related Openclaw Skills
LinkedIn Automation
A professional growth engine for LinkedIn that automates content publishing, network expansion, and B2B lead acquisition.
Invoice Generator
A robust automation tool for generating professional, tax-compliant PDF invoices from structured data.
Rey Developer
A comprehensive autonomous coding framework designed to ensure code quality, security, and efficiency across multiple programming environments.
Deep Research
A sophisticated investigation engine that performs multi-source data gathering and cross-platform analysis to generate structured professional reports.
Intelligent Model Router
An automated model routing skill that dynamically switches between Claude Sonnet 4.5 and Opus 4.6 based on task complexity to balance cost and quality.
Nano Banana Pro: AI Image Generation
A professional-grade image generation skill powered by Gemini that excels at rendering Japanese text and creating platform-optimized visuals.
Featured*
