Keep Learning is an intelligent synchronization skill that transforms your local Markdown files and source code into a structured, searchable knowledge base within your AI agent's memory.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install keep-learning
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 keep-learning 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 Keep Learning?
Keep Learning is a sophisticated utility designed for the Openclaw Skills ecosystem that bridges the gap between your local documentation and your AI agent's active memory. It scans directories for Markdown and a wide variety of programming languages, extracting core concepts and decisions to build a high-performance, three-layer knowledge architecture. This ensures that your agent doesn't just read your files, but actually understands and recalls the expert experience and project overviews contained within them.
By leveraging the Openclaw Skills framework, Keep Learning implements an incremental learning engine. It intelligently detects changes via Git diffs or local file monitoring, ensuring your AI's internal knowledge stays synchronized with your local edits without redundant processing. This makes it an indispensable tool for developers and researchers who need their AI collaborators to stay current with evolving project documentation and local codebases.
Keep Learning Use Cases
- Synchronizing team documentation from a local Git repository into an AI agent's long-term memory.
- Onboarding an AI agent to a new codebase by indexing source files across Python, JavaScript, Java, and Go.
- Transforming a personal Obsidian or Markdown-based second brain into an actionable knowledge base for Openclaw Skills.
- Maintaining an updated index of project-specific best practices and expert experience for real-time AI assistance.
How Keep Learning Works
- Configuration Retrieval: The skill searches memory for an existing knowledge base path or prompts the user to provide one if it is the first run.
- Repository Update: If a Git repository is detected, it performs a git pull to ensure the most recent data is available.
- Incremental Change Detection: It compares the current state against the last-commit hash and scans uncommitted changes to identify only the files that need re-processing.
- Knowledge Extraction: The skill parses supported file formats to identify themes, core concepts, and key conclusions.
- Multi-Layer Storage: It populates L1 Core Memory for instant recall and L2 Knowledge Indexes for file-path mapping, while leaving L3 source files on the local disk for deep-dives.
- Deduplication and Reporting: It checks for existing memory entries to prevent duplicates and generates a comprehensive learning report for the user.
Keep Learning Setup
To use this within the Openclaw Skills environment, ensure your agent has filesystem access and follow these steps:
# Trigger the learning process directly via chat
keep learning
# Or use the specific Chinese trigger
持续学习知识
When prompted, provide the absolute path to your documentation folder (e.g., ~/docs/my-project). Ensure git is installed if you plan to use the incremental update features.
Keep Learning Data Schema & Taxonomy
Keep Learning organizes data into a tiered hierarchy to balance performance and depth:
| Layer | Storage Type | Content Type | SEO Value |
|---|---|---|---|
| L1: Core Memory | Agent Memory | Key conclusions and best practices | Instant recall |
| L2: Index | Agent Memory | File paths, summaries, and keywords | Searchability |
| L3: Source | Local Filesystem | Full original content and code | Deep-dive context |
All runtime metadata, including configuration and the last-processed Git hash, is stored locally in ~/.keep-learning/.
Keep Learning Advanced Features
- Incremental Learning Engine: Specifically optimized for Openclaw Skills to only process files that have changed since the last session.
- Git Integration: Automatically pulls latest updates and detects both committed and uncommitted local changes.
- Intelligent Deduplication: Uses search_memory before creation to update existing knowledge nodes rather than creating clutter.
- Multi-Category Classification: Automatically sorts knowledge into expert_experience, project_introduction, or learned_skill_experience based on content analysis.
SKILL.md
Loading
Related Openclaw Skills
Aight Utils
Aight Utils provides a native bridge between your AI agent and the Aight iOS app to manage reminders, tasks, and deadlines in real-time.
3D Web Experience
An expert framework for building immersive 3D web experiences using Three.js, React Three Fiber, and Spline.
Cert Decode
A powerful utility for parsing and inspecting X.509 PEM certificates to extract human-readable SSL/TLS metadata.
Home Layout
A comprehensive interior design and spatial optimization tool that provides intelligent layout solutions and furnishing inspiration.
Snipcart
A comprehensive e-commerce integration for managing Snipcart store data, products, and customer lifecycles via the Membrane CLI.
Syncro
A comprehensive integration for Syncro that enables AI agents to manage MSP operations, ticketing, and remote monitoring through a unified interface.
Featured*
