What is the difference between the L1 and L2 layers in this system?

The L1 Context Layer contains raw, daily situational logs, while the L2 Behavior Layer stores verified, long-term habits and preferences extracted from those logs.

What happens if no patterns are found during the weekly review?

If no behaviors meet the '3+ occurrences' threshold, the skill will inform you that no obvious patterns were found and suggest continued observation.

Why does the skill require mem-file-scan as a prerequisite?

Running a file scan ensures that edits made directly in external tools like Obsidian are captured, providing a complete picture of your week beyond just chat history.

Can I edit a suggested behavior before it is saved to long-term memory?

Yes, the skill presents a candidate list where you can choose to record the pattern as-is, ignore it, or modify the description before it is finalized in the L2 layer.

Memory Weekly Review

Name: Memory Weekly Review
Author: zephyrwang6

byzephyrwang6

•

生産性とワークフロー

Automates the transition of daily context records into long-term behavioral patterns within an AI personal memory system.

The Memory Weekly Review (mem-weekly) skill is a specialized component for the AI Personal Memory System designed to refine short-term situational observations into long-term behavioral insights. It systematically analyzes weekly logs and file modifications to identify recurring habits, tool preferences, and decision-making patterns that appear three or more times. By presenting these findings as a candidate list for user confirmation, it ensures that your AI assistant evolves accurately based on your actual workflows, promoting verified data from the L1 Context Layer to the L2 Behavior Layer while maintaining a detailed historical record of your weekly progress.

主な機能

01Interactive Candidate Lists: Generates a clear overview of findings for user approval, rejection, or modification.

02Intelligent Pattern Recognition: Identifies habits or tool preferences appearing 3+ times per week.

0337 GitHub stars

04Multi-Layer Memory Promotion: Seamlessly migrates verified data from the Context Layer to the permanent Behavior Layer.

05Automated L1 Context Analysis: Scans weekly logs and file modifications to extract relevant situational events.

06Comprehensive Meta-Logging: Automatically updates a central review ledger to track key events and behavioral shifts over time.

ユースケース

01Tracking and optimizing decision-making patterns by analyzing historical situational records.

02Refining AI personal assistance by documenting project management preferences and communication styles.

03Codifying recurring coding habits and preferred tech stacks based on actual weekly activity.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zephyrwang6/myskill mem-weekly

For use in Claude.ai and ChatGPT

Download Skill