Surrealdb Knowledge Graph Memory

What This Skill Does

SurrealDB Knowledge Graph Memory is a robust, self-improving memory architecture for OpenClaw agents. It transitions your agent from stateless interactions to a persistent, evolving intelligence. By leveraging a local SurrealDB instance, it stores facts as structured nodes and edges, enabling the agent to recall past experiences, understand complex relationships between data, and learn from its own history. The system operates via a continuous feedback loop: it records memories in markdown files, parses them via LLMs to extract structured data, correlates entities to build a knowledge graph, and injects context into future conversations to ensure the agent grows smarter over time.

Installation

To install this skill, run: clawhub install openclaw/skills/skills/maverick-software/surrealdb-knowledge-graph-memory. Ensure your environment has python3 (v3.10+) and the surreal binary installed in your system path. You must also configure your OPENAI_API_KEY in the environment variables, as the extraction logic uses OpenAI models to parse unstructured text into graph data.

Use Cases

• Project Management: Track long-running task states, client requirements, and historical technical debt across multiple sessions.
• Personalized Assistance: Maintain a profile of user preferences, past topics discussed, and specific context relevant to the user's workflow.
• Autonomous Research: Build a research assistant that connects dots between disparate sources, forming a web of knowledge rather than just storing isolated documents.
• System Automation: Use the episodic memory to audit agent actions and refine future decision-making based on prior task success or failure.

Example Prompts

1. "Remember this: The project requirements for the new API integration prioritize security-first authentication protocols."
2. "What do you know about the authentication bugs we discussed last Tuesday in the legacy module?"
3. "Memory search: Find similar tasks where we faced issues with dependency hell during the Python 3.10 migration."

Tips & Limitations

• Privacy: Since this skill utilizes external LLMs for entity extraction, ensure no sensitive personal information is stored in plain-text markdown files if your security policy forbids it.
• Performance: The graph grows over time; for extremely large workspaces, ensure your SurrealDB instance is optimized for memory consumption.
• Dependencies: The 6-hour extraction cycle and 3 AM correlation job require your host machine to be running at those times. If the machine is off, the jobs will trigger upon the next startup.