triple-memory-baidu-embedding

What This Skill Does

The triple-memory-baidu-embedding skill provides a robust, three-tiered memory architecture for OpenClaw AI agents. By combining Baidu Embedding-V1 for vector-based semantic retrieval, Git-Notes for branch-specific structured documentation, and local file-based workspace search, it ensures the agent maintains a cohesive context across long-running sessions. The architecture is specifically optimized for Chinese language processing and provides a privacy-conscious way to store and recall user preferences, project decisions, and task statuses without relying solely on global cloud memory solutions. It acts as an "external brain" for your agent, allowing it to remember what you previously decided, what your specific preferences are, and where your project files are located.

Installation

1. Ensure your environment has the ClawHub CLI installed.
2. Execute clawdhub install openclaw/skills/skills/xqicxx/triple-memory-baidu-embedding to fetch the core modules.
3. Install the required supporting packages: clawdhub install git-notes-memory and clawdhub install memory-baidu-embedding-db.
4. Configure the Baidu API by adding your credentials to the system environment: BAIDU_API_STRING and BAIDU_SECRET_KEY.
5. Copy the file search utility via scripts/file-search.sh to your local workspace directory to enable full file-system awareness.

Use Cases

• Project Continuity: Retain technical decisions and architectural choices across multiple coding sessions without needing to re-summarize past progress.
• Personalization: Store user-specific style preferences, coding standards, or project constraints that the agent will automatically recall in future interactions.
• Knowledge Management: Use Git-Notes to track "critical" or "high" priority decisions that need to be linked to your git commit history, ensuring documentation evolves alongside the codebase.
• Multilingual Contexts: Leverage the Baidu Embedding integration for superior semantic search capabilities when working with documents written primarily in Chinese.

Example Prompts

1. "I prefer using Tailwind CSS for all front-end styling in this project; remember this for future UI components."
2. "What was the reasoning behind our decision to switch to a PostgreSQL database last week? Check the memory and summarize."
3. "Save this technical debt task as a high-priority memory tagged with 'refactoring' and 'legacy-code'."

Tips & Limitations

• Initialization: Always run the sync command python3 skills/git-notes-memory/memory.py -p $WORKSPACE sync --start at the beginning of every session to ensure memory pointers are current.
• API Dependency: The Baidu Embedding layer is disabled if API credentials are not valid; the system will fall back to basic file searching if the API service is unreachable.
• Privacy: While Git-Notes is purely local, the Baidu Embedding component transmits metadata to the Baidu API for vectorization. Ensure sensitive internal data is scrubbed if you have strict external compliance requirements.