Lobster Memory Master is a comprehensive long-term memory management system that transforms stateless AI into a stateful assistant with persistent recall and automated context compaction.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install lobster-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 lobster-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 Lobster Memory Master?
Lobster Memory Master is a sophisticated architectural framework designed to solve the problem of AI forgetfulness. By implementing a structured memory hierarchy, this Openclaw Skills contribution allows AI agents to maintain persistent knowledge of user preferences, past decisions, and project milestones across multiple sessions.
The system utilizes a multi-tier storage approach, including a high-velocity Working Buffer for active tasks and an archival system for historical data. It ensures that your AI remains performant by automatically compacting and archiving data when context usage exceeds efficiency thresholds, making it an essential tool for developers building autonomous daily assistants.
Lobster Memory Master Use Cases
- Managing long-term recall of user preferences, past project decisions, and key dates.
- Automatically compacting and archiving context when LLM token limits reach 60% capacity.
- Establishing an autonomous daily learning cycle for AI agents to improve over time.
- Creating a persistent state for AI assistants across different sessions or IDE restarts.
How Lobster Memory Master Works
- The AI monitors context usage via the Heartbeat Protocol to ensure efficiency.
- When memory thresholds are met, the skill triggers the Working Buffer protocol to extract key dialogue fragments.
- Data is systematically categorized and written into MEMORY.md for long-term storage and session-specific files for immediate state.
- Scheduled cron jobs execute maintenance tasks like Daily Skill Discovery and nightly archiving to memory/YYYY-MM-DD.md files.
- During retrieval, the agent performs a coordinated search across the memory directory to answer user queries with precise citations.
Lobster Memory Master Setup
To initialize the memory architecture for your agent, execute the following commands in your workspace:
# Create the essential long-term memory file structure
touch MEMORY.md
mkdir -p memory
touch memory/working-buffer.md
touch SESSION-STATE.md
touch HEARTBEAT.md
After creating the files, populate the MEMORY.md file with your initial user profile and project context to ground the agent.
Lobster Memory Master Data Schema & Taxonomy
Lobster Memory Master organizes data through a clear file-based taxonomy to ensure interoperability within Openclaw Skills:
| File/Folder | Purpose | Update Frequency |
|---|---|---|
| MEMORY.md | Primary long-term storage for preferences and decisions | Weekly Distillation |
| SESSION-STATE.md | Tracks current session objectives and metadata | Real-time |
| memory/ | Directory containing daily snapshots (YYYY-MM-DD.md) | Daily |
| working-buffer.md | Temporary storage for high-context active segments | Triggered at 60% Context |
| learning-log.md | Records new skills acquired and project progress | Daily |
Lobster Memory Master Advanced Features
- Heartbeat Protocol: Automated self-checks for context utilization and memory health.
- Context Compaction: Intelligent archiving that prevents token overflow while preserving critical decision history.
- Autonomous Cron Integration: Support for scheduled daily tasks such as memory freshening and skill discovery.
- Citation-Based Retrieval: Sophisticated memory search that provides exact file and line references for transparency.
SKILL.md
Loading
Related Openclaw Skills
Smart Answering Framework
A sophisticated reasoning framework that dynamically adapts response structures based on question types to deliver evidence-based, natural-sounding answers.
Python Code Review
An expert AI reviewer that enforces PEP8 standards, type safety, and efficient async patterns in Python codebases.
Construction Safety Checklist
A conversational AI tool that guides site supervisors through a 10-category safety inspection to generate professional HSE reports with automated severity classification.
Social Growth Ops
A professional-grade content operations engine for planning, drafting, and safely automating multi-platform social media growth.
Sora 2 Generate
A professional-grade skill for generating high-fidelity AI videos from text prompts or public image URLs using the Sora 2 engine.
Follow-Through Day (FTD) Detector
A market timing state machine that detects Follow-Through Day signals to confirm market bottoms using the S&P 500 and NASDAQ indices.
Featured*
