An agent-controlled memory plugin that provides persistent, file-based knowledge storage with semantic search for Openclaw Skills.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-tools
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-tools 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-tools?
memory-tools is a robust persistence layer designed for the modern AI workflow within Openclaw Skills. Unlike traditional memory systems that automatically capture every interaction and flood the context window with noise, this plugin follows the AgeMem approach where the agent itself decides when to store, update, or retrieve information. This leads to higher precision and more relevant context during long-running interactions.
Version 2 of this skill transitions to a fully local, file-based storage system using Markdown and YAML. This ensures that your data remains human-readable and portable, requiring no external APIs or OpenAI dependencies. By keeping everything on-disk, memory-tools provides a secure and transparent way to manage agent knowledge while supporting advanced features like local vector search through QMD.
memory-tools Use Cases
- Storing and recalling user preferences across different chat sessions to personalize experiences.
- Managing long-term project context and technical decisions in a searchable format.
- Maintaining standing instructions and personality traits that the agent can auto-inject.
- Tracking relationship data and entity facts without cluttering the immediate conversation history.
- Handling temporal events and schedules that require expiration or decay logic.
How memory-tools Works
- The AI agent identifies a piece of information worth remembering or a need to recall past data.
- The agent invokes one of the specialized memory tools such as memory_store or memory_search.
- The plugin processes the request, either writing a new Markdown file with YAML frontmatter to the local storage or querying existing files.
- If QMD is installed, the search process utilizes BM25 and vector reranking via local GGUF models for high-accuracy retrieval.
- The retrieved memories are presented back to the agent, allowing it to incorporate historical context into its current response.
memory-tools Setup
To integrate this capability into your Openclaw Skills environment, follow these steps:
# Install the package
clawhub install memory-tools
# Navigate to the skill directory and build
cd skills/memory-tools
npm install
npm run build
# Link and enable the plugin
openclaw plugins install --link .
openclaw plugins enable memory-tools
# Restart your gateway to apply changes
openclaw gateway restart
For enhanced semantic search, optionally install QMD:
npm install -g @tobilu/qmd
memory-tools Data Schema & Taxonomy
Memories are organized into specific categories and stored as individual Markdown files within the ~/.openclaw/memories/ directory. This structure makes Openclaw Skills data easy to audit.
| Category | Purpose | Example |
|---|---|---|
| fact | Static data | User's birthday |
| preference | User likes/dislikes | Dark mode preference |
| instruction | Standing orders | Always use TypeScript |
| decision | Past choices | Used PostgreSQL for the DB |
Each file contains a YAML frontmatter block with fields for id, confidence, importance, created_at, and tags, followed by the raw text content.
memory-tools Advanced Features
- Confidence and Importance Scoring: Allows agents to weigh the reliability and priority of different memories.
- Automatic Decay: Temporal memories can be configured to expire or become stale over time.
- Local Semantic Search: Integration with QMD enables vector search and BM25 ranking without sending data to third-party providers.
- Auto-Injection: The ability to automatically prepend critical standing instructions at the start of every agent session.
- Seamless Migration: Built-in tools to upgrade legacy SQLite or LanceDB data from v1 to the new file-based format.
SKILL.md
Loading
Related Openclaw Skills
Kalshi
The Kalshi skill for Openclaw Skills enables automated trading and market analysis on the first CFTC-regulated prediction market exchange.
Xobni.ai Email Skill
Provide AI agents with real-world email addresses, semantic search capabilities, and MCP server integration for seamless communication.
agentcli-go
A comprehensive Go framework for scaffolding and building standardized CLI tools optimized for AI agent integration.
PostalForm Machine Order
A professional workflow for AI agents to autonomously send physical letters and documents via PostalForm using secure machine payments.
Tick Multi-Agent Coordination
A Git-backed Markdown protocol that provides a transparent, machine-parseable framework for coordinating complex tasks between humans and AI agents.
Switchboard
Switchboard optimizes AI agent operational costs by intelligently routing tasks to different model tiers based on their complexity and requirements.
METADATA
Featured*
