A technical framework for configuring vector-based long-term memory and persistent context for AI coding agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-setup
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-setup 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 Setup Skill?
The Memory Setup skill is a comprehensive configuration framework designed to eliminate context loss in AI-driven development workflows. By leveraging Openclaw Skills, developers can transform their agents from stateless participants into context-aware assistants that remember past decisions, project history, and specific user preferences across multiple sessions.
This skill establishes a structured environment for storing and retrieving information using semantic search. It integrates with embedding providers to index daily logs, session transcripts, and a curated long-term memory file. This ensures that the agent can proactively recall relevant technical details, reducing repetitive questioning and improving the continuity of complex software projects.
Memory Setup Skill Use Cases
- Maintaining continuity in long-term software engineering projects where context spans several weeks.
- Retrieving specific architectural decisions or lessons learned from past coding sessions.
- Managing personalized developer preferences, including coding styles and specific tool configurations.
- Automatically indexing daily logs to track progress and pending follow-up tasks.
- Creating a shared knowledge base for multi-agent workflows within a unified workspace.
How Memory Setup Skill Works
- The configuration is defined in the global settings to enable the memory search engine and select an embedding provider.
- A dedicated directory structure is initialized in the workspace to separate logs, project context, and system preferences.
- A central MEMORY.md file is created to act as the primary source of truth for high-level project facts and user context.
- The agent uses the memory search functionality to perform vector-based queries against the indexed Markdown files.
- Context is dynamically injected into the conversation based on the semantic relevance of the retrieved snippets.
Memory Setup Skill Setup
To enable persistent context via Openclaw Skills, add the following configuration to your ~/.clawdbot/clawdbot.json or moltbot.json file:
{
"memorySearch": {
"enabled": true,
"provider": "voyage",
"sources": ["memory", "sessions"],
"indexMode": "hot",
"minScore": 0.3,
"maxResults": 20
}
}
Initialize your workspace directory structure with these commands:
mkdir -p memory/logs memory/projects memory/groups memory/system
touch MEMORY.md
Finally, ensure your API keys for providers like Voyage AI or OpenAI are exported in your environment or added to your agent's secret management.
Memory Setup Skill Data Schema & Taxonomy
The skill utilizes a hierarchical Markdown schema to organize and index information for efficient retrieval.
| Data Component | Location | Content Type |
|---|---|---|
| Long-term Memory | /MEMORY.md |
Curated facts, project summaries, and decisions |
| Daily Activity | /memory/logs/ |
Date-stamped logs (YYYY-MM-DD.md) recording daily tasks |
| Project Context | /memory/projects/ |
Detailed technical documentation and project-specific notes |
| User Preferences | /memory/system/ |
Development environment setups and communication styles |
| Session History | /sessions/ |
Automatically generated conversation transcripts |
Memory Setup Skill Advanced Features
- Support for Hot Mode indexing which provides real-time updates to the vector database as you work.
- Multi-provider flexibility allowing the use of Voyage AI, OpenAI, or local embedding models for offline environments.
- Granular relevance control through adjustable minScore thresholds to tune search precision.
- Automated daily log formatting to streamline the capture of technical follow-ups.
- Integrated agent instructions for standardized memory retrieval protocols.
SKILL.md
Loading
Related Openclaw Skills
Cron Retry
An automated recovery system that detects and restarts cron jobs failed by network interruptions.
JRB Remote Site API
A comprehensive WordPress management skill for AI agents to control content, themes, and CRM integrations via a secure REST API.
Voku Language
Voku is a precision-engineered constructed language designed for zero-ambiguity communication between AI agents.
Download Tools
A streamlined collection of command-line scripts for downloading WeChat articles as text and YouTube videos as high-quality audio files.
Flight Pricer
A professional-grade CLI tool for searching flight offers and pricing through the Duffel API.
soccer-cli
A high-performance Go-based command-line interface for fetching real-time football data and match analytics.
Featured*
