MemWeaver is a memory profiling engine that mines hidden behavioral patterns from an AI agent's logs to generate structured, confirmed user profiles.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memweaver
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memweaver using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is MemWeaver?
MemWeaver is a sophisticated memory analysis skill designed to bridge the gap between an AI's raw memory storage and true user understanding. Unlike simple retrieval tools, it functions as an insight mining engine that digs into long-term memory files and daily logs to uncover preferences, behavioral contradictions, and hidden traits that the user might not even recognize themselves.
By analyzing patterns in decision-making, energy allocation, and communication styles, MemWeaver moves beyond mere data storage. It facilitates a deeper alignment between the user and their AI agent, ensuring the agent's persona and assistance are perfectly calibrated to the user's actual behavior rather than just their stated preferences.
MemWeaver Use Cases
- Discovering discrepancies between your stated goals and actual time allocation in daily logs.
- Generating a structured YAML profile to synchronize your preferences across different AI agents or platforms.
- Periodically re-evaluating project priorities based on real engagement levels recorded in memory.
- Identifying unlabeled skills and professional interests that emerge frequently in work logs.
- Refining AI communication styles by analyzing historical response depth and format preferences.
How MemWeaver Works
- The system collects data by reading MEMORY.md and recent daily logs via a Python script, calculating token counts for analysis.
- An LLM performs a deep analysis across three dimensions: basic identity extraction, hidden pattern mining (identifying contradictions), and project importance re-evaluation.
- The skill initiates an interactive, batch-mode questionnaire (usually 10-15 questions) where the user confirms or clarifies the AI's findings.
- Based on user feedback, the AI updates its confidence scores and refines the analysis.
- A final, structured user profile is generated in YAML format and saved to the output directory for future use by Openclaw Skills.
MemWeaver Setup
MemWeaver requires Python 3.8+ and no external dependencies. To start the profiling process, use the following commands:
# Navigate to your agent directory and collect the last 14 days of memory
cd {baseDir} && python3 scripts/collect_memory.py --days 14
# After analysis and interactive quiz, the profile is saved
cd {baseDir} && python3 scripts/save_profile.py --file /tmp/memweaver_profile.yaml
MemWeaver Data Schema & Taxonomy
The skill organizes data into a structured hierarchy focusing on behavioral metadata and project statuses:
| Category | Data Points Included |
|---|---|
| Identity | Role, Tech Stack, MBTI, Confidence Scores |
| Work Patterns | Decision Style, Creation Preference, Energy Sources |
| Interests | Professional and Personal Topics with depth/importance levels |
| Goals | Career and Product goals with priority and timeframe |
| Hidden Patterns | Evidence-based findings and user confirmation status |
| Projects | Comprehensive list of active, shelved, or archived projects |
MemWeaver Advanced Features
- Hidden Insight Mining: Specifically targets contradictions between user statements and behavioral records.
- Batch Interactive Mode: Uses personality-test style batching (5 questions at a time) to minimize user friction.
- Confidence Scoring: Every extracted trait is assigned a confidence value based on the source (explicit vs. inferred).
- Automated Backups: The profile saving script automatically archives older versions before generating the new output/profile_YYYYMMDD.yaml file.
- Scenario Reconstruction: Generates Type A questions that reconstruct real scenes from memory to validate factual accuracy.
SKILL.md
Loading
Related Openclaw Skills
EternalClaw Memory Sync
A secure synchronization utility for Openclaw agents to restore memory states from remote URLs using military-grade encryption.
cuihua-logger
An AI-powered logging assistant that generates production-ready structured logs with smart level selection and performance tracking.
Calendar Sync
A comprehensive calendar synchronization tool for managing Google Calendar, Outlook, and local events via a unified interface.
Slide to Video Converter
An end-to-end automated pipeline for transforming presentation slides into professional narrated videos with synchronized subtitles.
Video Content Watcher
An AI-powered skill that extracts audio, transcribes speech, and generates comprehensive Markdown reports from YouTube, Bilibili, and local videos.
Vercel Staging Workflow
A professional automation skill that establishes stable, auto-updating staging URLs for Vercel projects using GitHub Actions.
Featured*
