triple-memory

What This Skill Does

The Triple Memory System is an advanced, multi-layered memory architecture designed to provide OpenClaw agents with comprehensive, persistent context retention. Rather than relying on a single data store, it intelligently orchestrates three distinct memory backends: LanceDB for fluid, auto-recall conversation history; Git-Notes for structured, branch-aware decision tracking; and Workspace File Search for grounding the agent in the physical documentation and codebase state. This synergistic approach allows the agent to handle ephemeral chat contexts, long-term project decisions, and static file data simultaneously, ensuring that critical user preferences are never lost between sessions.

Installation

To integrate this system, start by installing the core components via the OpenClaw plugin ecosystem. First, run 'clawhub install openclaw/skills/skills/ktpriyatham/triple-memory'. Next, configure the LanceDB plugin within your configuration file by mapping the 'memory' slot to 'memory-lancedb' and providing a valid OpenAI API key for embedding generation. Ensure that 'autoRecall' and 'autoCapture' are set to true to enable the agent's subconscious memory functions. Finally, ensure the 'scripts/file-search.sh' utility is placed in your project root, and initialize the Git-Notes system using 'python3 skills/git-notes-memory/memory.py -p $WORKSPACE sync --start' at the beginning of every session.

Use Cases

Use this skill when your development project requires high-level consistency, such as architectural decision tracking or complex task management across multiple git branches. It is ideal for teams needing an agent that remembers technical constraints, stylistic preferences, and project-specific documentation updates. It is particularly effective for deep-dive coding sessions where you need to reference previous decisions without manually restating them.

Example Prompts

1. "Remember that we are standardizing all API responses to follow the JSON:API specification; treat this as a critical project requirement."
2. "Search the workspace for any existing database connection strings and summarize our current configuration status."
3. "Recalling our conversation from last week, what was the reason we decided to switch from Redis to Memcached for the caching layer?"

Tips & Limitations

To maximize effectiveness, use explicit importance flags when interacting with Git-Notes. Labeling decisions as 'critical' or 'high' importance ensures they appear prominently during future queries. Note that this system does require a local git repository to function at peak capacity. While LanceDB handles conversation history automatically, maintain your workspace documentation files regularly, as the file search relies on the accuracy of your markdown files. Do not treat memory as a replacement for long-term backups; always commit your configuration changes to your main repository.