smart-memory-lite-automaton

What This Skill Does

The Smart Memory Lite Automaton skill provides a lightweight, persistent cognitive layer for your AI agents. By default, most AI sessions are stateless, meaning the agent "forgets" everything once a chat concludes. This skill integrates directly into your workflow to capture, organize, and recall conversation history, preferences, and session-specific data. It leverages a file-based storage architecture, ensuring that your data remains human-readable, portable, and easily backed up without needing complex external database infrastructures or volatile API dependencies.

Installation

To integrate this memory system into your agent project, execute the following command in your terminal:

clawhub install openclaw/skills/skills/chenghaifeng08-creator/smart-memory-lite-automaton

Once installed, initialize the SmartMemory class within your agent initialization logic, providing a persistent path for the JSON storage files. The setup is designed for zero-configuration, allowing you to begin saving interactions immediately.

Use Cases

1. Long-term Persona Maintenance: Ensure the agent remembers user preferences (e.g., "always use Python" or "keep responses concise") across different calendar days.
2. Context-Aware Research: Retrieve specific insights or findings from past brainstorming sessions, preventing the repetition of information.
3. Session Continuity: Allow agents to reference previous project statuses or open topics when a user returns to a conversation after an extended break.
4. Personalized Coaching: Track progress or recurring themes in user requests to provide better, more tailored advice over time.

Example Prompts

1. "What was the key decision we reached about the database schema during our conversation on Tuesday?"
2. "Review my previous interaction history regarding crypto and list the top three topics I have shown interest in."
3. "Summarize my last seven days of activity to provide a status update on the current project requirements."

Tips & Limitations

• Tip: Utilize the tagging system when saving memories to make the recallByTag function more effective. Tags like 'urgent', 'documentation', or 'code-snippets' allow for much faster retrieval.
• Tip: Review the generated JSON files in your storage directory periodically. Because they are plain text, you can manually clean up or organize specific memory blocks if necessary.
• Limitation: As this is a 'Lite' system, it does not support semantic vector searching natively. It relies on direct keyword and tag matching, which is highly efficient but lacks the fuzzy logic of a dedicated Vector DB.
• Limitation: Memory usage scales with the number of conversations. For very large datasets, consider archiving older JSON files to maintain high performance.