A professional-grade local memory architecture for AI agents utilizing a three-tier system to prevent context loss and optimize knowledge retrieval.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-manager-clawd
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-manager-clawd 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 Manager?
Memory Manager is a sophisticated local memory management solution designed specifically for AI agents within the Openclaw Skills ecosystem. By implementing a structured semantic, procedural, and episodic memory pattern, it moves beyond inefficient flat-file storage to a system that improves retrieval accuracy by up to 18.5%. It ensures that agents never lose critical context by monitoring compression risks and providing automated snapshotting.
This skill provides a human-readable, 100% local architecture that requires no external dependencies or API costs. It allows agents to distinguish between what happened (episodic), what they know (semantic), and how to perform tasks (procedural), resulting in a more grounded and capable digital identity.
Memory Manager Use Cases
- Detecting memory compression risks before critical context is lost during long agent sessions.
- Organizing flat memory logs into structured knowledge bases for better long-term recall.
- Searching historical agent actions and events using time-based episodic retrieval.
- Documenting and retrieving complex step-by-step workflows through procedural memory.
- Maintaining data privacy by keeping all agent memories and identity markers on local hardware.
How Memory Manager Works
- The system initializes a dedicated directory structure consisting of episodic, semantic, procedural, and snapshot folders.
- Agents log raw chronological events into the episodic memory tier on a daily basis.
- The organization engine parses these logs to extract distilled facts for semantic memory and reusable workflows for procedural memory.
- A detection script constantly monitors file sizes and context usage to provide safety alerts (Safe, Warning, or Critical).
- When thresholds are met, the system triggers automated snapshots to preserve state before memory pruning or compression occurs.
- High-speed local scripts allow for targeted searching across specific memory tiers or the entire knowledge base.
Memory Manager Setup
Installation
Initialize the structured memory environment using the provided setup script:
~/.openclaw/skills/memory-manager/init.sh
Integration
To automate memory health checks, add the following to your agent's heartbeat or crontab:
# Check for compression risks every 2 hours
~/.openclaw/skills/memory-manager/detect.sh
# Run daily organization and migration at 23:00
~/.openclaw/skills/memory-manager/organize.sh
Memory Manager Data Schema & Taxonomy
The skill organizes data into a hierarchical file-based taxonomy for maximum transparency:
| Directory | Memory Type | Description |
|---|---|---|
memory/episodic/ |
Episodic | Chronological daily event logs (e.g., YYYY-MM-DD.md) |
memory/semantic/ |
Semantic | Distilled facts, concepts, and topical knowledge |
memory/procedural/ |
Procedural | Step-by-step guides and reusable process workflows |
memory/snapshots/ |
Snapshots | Backups generated during critical compression events |
memory/legacy/ |
Legacy | Original flat files preserved after migration |
Memory Manager Advanced Features
- Multi-tier memory categorization to separate events from reusable logic.
- Automated compression detection with visual status indicators (Safe/Warning/Critical).
- Semantic search capabilities across different memory types using optimized local scripts.
- Manual categorization overrides via CLI for precise knowledge management.
- Migration tools to safely transition legacy flat-file memories into the structured Openclaw Skills format.
- Detailed usage statistics to track memory growth rates and compression event history.
SKILL.md
Loading
Related Openclaw Skills
Markdown to HTML Converter
A professional utility that converts Markdown files into high-quality HTML pages featuring interactive table of contents, LaTeX support, and offline capabilities.
Fun-Denoise
Fun-Denoise is a professional-grade AI audio noise reduction service powered by Alibaba Tongyi Lab algorithms to isolate clean vocals from background noise.
English Daily Report
A powerful language learning automation that generates comprehensive daily English news reports with translations, audio, and PDF formats.
Junyi Client Followup
A specialized automation tool for collecting monthly growth feedback from parents and streamlining it for professional planners.
GeoJSON Utility
A robust diagnostic tool for inspecting, validating, and troubleshooting GeoJSON files and payloads to ensure spatial data integrity.
A2A Market
A2A Market is a decentralized exchange for AI agent capabilities, enabling agents to autonomously buy, sell, and monetize skills using USDC or credits.
Featured*
