A sophisticated memory management system for AI agents that mimics human brain mechanisms like neuroplasticity, myelination, and tiered long-term storage.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install brain-evolution-skill
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 brain-evolution-skill 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 brain-evolution?
The brain-evolution skill is a revolutionary intelligence management system designed for the Openclaw Skills framework. Inspired by human evolutionary biology, it utilizes mechanisms such as synaptic pruning and the Ebbinghaus forgetting curve to manage AI memory efficiently. By implementing a layered architecture (L0/L1/L2), it ensures that high-frequency tasks become 'conditioned reflexes' with near-zero latency, while rarely used data is naturally phased out to save resources.
Version 2.0 represents a major leap forward, integrating high-performance SQLite storage with Write-Ahead Logging (WAL) and thread-safe concurrency. This allows multiple agents to share a unified, evolving 'brain' that prioritizes important information and optimizes neural pathways based on real-world usage patterns, significantly reducing the computational overhead of redundant LLM calls.
brain-evolution Use Cases
- Optimizing AI agent dialogue by forming reflexes for common user interactions.
- Accelerating high-frequency tool calls to achieve response times as low as 0.01ms.
- Personalized learning that adapts autonomously to specific user habits and preferences.
- Automatic memory maintenance to prune low-frequency knowledge and prevent database bloat.
How brain-evolution Works
- Reflex Check (L0): The system first checks the Reflex Layer for high-strength, high-frequency responses; if matched, it returns instantly (0.01ms).
- Smart Cache (L1): If L0 misses, it queries the Cache Layer, utilizing both memory LRU and SQLite disk storage (0.1ms - 10ms).
- Neural Pathway (L2): If L1 misses, the system searches the long-term Pathway Layer which uses indexed SQLite lookups (100ms).
- Task Execution & Learning: If no memory exists, the agent performs the actual task and calls the
recordmethod to initiate the learning process. - Reinforcement & Decay: Every interaction strengthens the neural pathway (Myelination), while inactivity leads to gradual strength reduction based on the forgetting curve.
brain-evolution Setup
To integrate brain-evolution into your project, ensure it is placed in your skills/ directory and use the following commands for management:
# Navigate to the skill directory
cd skills/brain-evolution
# Check the current status of the brain (reflexes, pathways, etc.)
python3 brain_manager.py status
# Run the v2.0 performance and logic tests
python3 brain_evolution_v2.py
# Manually trigger the forgetting mechanism to clean up weak memories
python3 brain_manager.py forget
brain-evolution Data Schema & Taxonomy
The skill organizes data within the brain_data/ directory using a multi-tiered file structure for maximum efficiency:
| File Name | Description | Format |
|---|---|---|
neural_pathways.json |
The core network of long-term memory pathways. | JSON |
reflex_cache.json |
Stores high-speed conditioned reflexes for L0 access. | JSON |
memory_trace.jsonl |
Append-only log of every learning event and access. | JSONL |
brain_report.json |
Summarized health and performance metrics of the system. | JSON |
brain-evolution Advanced Features
- Multi-Agent Concurrency: Full thread-safety via locking mechanisms, allowing parallel Openclaw Skills agents to access the same brain.
- Customizable Brain Parameters: Fine-tune the
REFLEX_THRESHOLD,DECAY_RATE, andFORGET_THRESHOLDto suit specific application needs. - Automatic Myelination: High-usage pathways automatically graduate to the reflex layer when strength exceeds 0.7.
- Tiered Latency Optimization: Massive performance gains with up to 200,000x acceleration for L0 reflex hits compared to raw execution.
SKILL.md
Loading
Related Openclaw Skills
Tianshu Mistake Sheet
Tianshu Mistake Sheet is a specialized utility for generating structured Markdown error logs and study trackers from raw text or templates.
Uplisting
A powerful integration to manage vacation rental properties, bookings, and guest communications through the Membrane ecosystem.
Sensitivity Charts
A powerful tool for visualizing how variable fluctuations impact project outcomes through automated tornado charts and scenario analysis.
Amazon Return Rate Optimizer
An intelligent analyzer designed to reduce Amazon return rates by diagnosing root causes, auditing listing accuracy, and streamlining FBA reimbursement claims.
SEO Audit Skill
An automated diagnostic tool designed to evaluate web pages for technical SEO integrity, content depth, and competitive performance.
Content Creation Pro
A comprehensive AI-driven content engine for multi-platform social media generation, scheduling, and trend analysis.
Featured*
