A three-tier memory architecture combining Baidu-powered vector recall, structured Git-Notes, and local workspace search.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install triple-memory-baidu-embedding
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 triple-memory-baidu-embedding 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 Triple Memory System with Baidu Embedding?
This skill implements a sophisticated triple-layer memory architecture designed for AI agents that require high context retention and privacy. By integrating Baidu Embedding for conversation auto-recall, Git-Notes for structured decision tracking, and file-based searching for workspace documentation, it ensures that your agent never loses track of user preferences or project history.
Developed as a specialized component of Openclaw Skills, this system focuses on privacy-first local storage and is specifically optimized for Chinese language semantics. It allows agents to function in a degraded mode using only local layers if API credentials are missing, ensuring robustness in various environments.
Triple Memory System with Baidu Embedding Use Cases
- Persistent context retention across multiple chat sessions for complex coding tasks.
- Privacy-conscious memory management using Baidu API with local vector storage.
- Structured decision tracking and entity extraction within Git-aware development environments.
- Optimizing AI agent interactions for superior Chinese language semantic understanding.
- Automatic preference and fact capture during natural conversation without manual triggers.
How Triple Memory System with Baidu Embedding Works
- The system automatically recalls relevant conversation memories via Baidu Embedding-V1 before generating an agent response.
- The agent synthesizes a response using context from all three layers: vector embeddings, structured Git-Notes, and local workspace files.
- Preferences, decisions, and facts are automatically captured and stored in the Baidu Embedding local vector database during the conversation.
- Specific technical decisions and entities are extracted and saved into the Git-Notes layer for branch-aware persistence.
- Workspace documents and logs are updated or searched using shell scripts to maintain long-term project documentation and daily logs.
Triple Memory System with Baidu Embedding Setup
Install the necessary components via the CLI using Openclaw Skills management tools:
clawdhub install git-notes-memory
clawdhub install memory-baidu-embedding-db
Configure your Baidu API credentials by setting the following environment variables:
export BAIDU_API_STRING='your_bce_v3_api_string'
export BAIDU_SECRET_KEY='your_secret_key'
Initialize the system within your workspace at the start of a session:
python3 skills/git-notes-memory/memory.py -p $WORKSPACE sync --start
Triple Memory System with Baidu Embedding Data Schema & Taxonomy
The skill organizes data across three distinct storage tiers to ensure both flexibility and persistence:
| Tier | Storage Mechanism | Data Type |
|---|---|---|
| Conversation | Baidu Embedding (Local Vector) | Unstructured dialogue history and preferences |
| Structured | Git-Notes | JSON objects with topics, entities, and importance levels |
| Workspace | Markdown Files | Curated MEMORY.md and daily logs in memory/*.md |
The Git-Notes layer specifically supports importance levels ranging from Critical (-i c) for explicit preferences to Low (-i l) for temporary notes.
Triple Memory System with Baidu Embedding Advanced Features
- Startup Hook Integration: Support for the memory-boot-loader hook to automatically run initialization scripts like session-init-triple-baidu.sh on gateway startup.
- Branch-Aware Memory: Keeps memories isolated per Git branch, ensuring that context from one feature branch doesn't interfere with another.
- Silent Operation: Executes all memory storage and retrieval tasks in the background without announcing actions to the user.
- Entity Extraction: Automatically identifies and indexes names, concepts, and topics within the Git-Notes layer.
- Degraded Mode Support: Gracefully falls back to Git-Notes and File Search if Baidu API credentials are not provided, maintaining core functionality.
SKILL.md
Loading
Related Openclaw Skills
Thinking Model Enhancer
An advanced cognitive framework that optimizes AI decision-making through multi-stage processing and integrated memory feedback loops.
FixClawd: AI System Repair & Troubleshooting Expert
A professional AI system repair expert that uses a strict priority-based workflow and advanced thinking models to diagnose and fix technical errors.
Memory Skills Sync
An automation utility designed to synchronize memory system skill files and scripts across directories with automated permission management.
Secure Memory Stack
A privacy-centric local memory system for AI agents that utilizes semantic search and versioned Git storage to maintain context without cloud dependency.
MuskInsider Pro
A minimal HTTP service providing Elon Musk news briefings and demonstration payment integration for AI agents.
OpenClaw Mem0 Plugin
A sophisticated memory integration that enables AI agents to remember user preferences, facts, and conversation history across multiple sessions.
Featured*
