Remember Me

What This Skill Does

The Remember Me skill, developed by achals-iglu, serves as the cognitive architecture for your OpenClaw agent. It enables the agent to maintain a coherent, persistent, and evolving understanding of your preferences, goals, and working style. Rather than starting every session from scratch, this skill allows the agent to build a durable profile of how you work, what you value, and what you need, ensuring that interactions become progressively more efficient and personalized over time. It utilizes a dual-tier system—daily ephemeral logs and a curated long-term memory file—to separate immediate context from foundational identity.

Installation

To add this capability to your agent, execute the following command in your OpenClaw terminal:

clawhub install openclaw/skills/skills/achals-iglu/remember-me

Ensure you have the necessary write permissions in your environment, as the skill requires creating and updating local documentation files (memory/YYYY-MM-DD.md and MEMORY.md) to function correctly.

Use Cases

• Project Management: Track ongoing goals and shift priorities across multiple coding or writing sessions without restating the context.
• Behavioral Personalization: Train your agent to adopt your preferred communication style, whether that means being hyper-concise or deeply explanatory.
• Boundary Enforcement: Set persistent privacy or workflow boundaries that the agent will respect consistently, such as "don't suggest external libraries" or "keep outputs under 200 words."
• Contextual Recall: Effortlessly retrieve decisions made in prior meetings or research phases.

Example Prompts

1. "Remember that for all my Python projects, I prefer using pytest over unittest and want docstrings in Google format."
2. "What did we decide were the key priorities for the Q4 dashboard update last week?"
3. "Stop suggesting UI mocks; I'm strictly focused on backend logic right now, so please keep our sessions focused on API design."

Tips & Limitations

• Be Explicit: The agent uses an integrity model (FACT, PREFERENCE, GOAL, HYPOTHESIS). If you want to ensure a setting persists, state it clearly as a fact or preference.
• Manage Your Memory: Occasionally review your MEMORY.md file. Since the agent can make hypotheses about your habits, clearing out outdated assumptions ensures the agent remains aligned with your current needs.
• Confidence Matters: If the agent is unsure, it will flag its reasoning as a 'hypothesis'. Don't be afraid to correct it; this is how the agent learns to distinguish between noise and actual intent.
• Scope: This skill is designed for contextual memory, not bulk data storage. Keep logs concise to maintain optimal retrieval speeds.