Mindkeeper bootstraps a production-grade, 3-layer memory system that enables Openclaw agents to remember, learn, and organize knowledge automatically.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install inkwell
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 inkwell 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 Mindkeeper?
Mindkeeper is a specialized framework designed to provide AI agents with a sophisticated, persistent memory architecture. By implementing Mindkeeper within Openclaw Skills, developers can transform a stateless agent into one that possesses a structured second brain. The system relies on a three-tier hierarchy: a core MEMORY.md for distilled tacit knowledge, daily session notes for raw capture, and a comprehensive PARA (Projects, Areas, Resources, Archive) structure for organized knowledge storage.
This skill ensures that knowledge is not just stored but is searchable and actionable. It integrates semantic search capabilities and automated maintenance routines to ensure the agent's context remains relevant and high-quality over time. Using Mindkeeper within the ecosystem of Openclaw Skills allows for a professional-grade knowledge management experience that mirrors human productivity workflows.
Mindkeeper Use Cases
- Setting up a permanent knowledge base for a new agent deployment.
- Organizing complex project data using the industry-standard PARA methodology.
- Implementing automated daily knowledge extraction from session transcripts.
- Enabling semantic and vector-based search across historical agent interactions.
- Establishing a formal end-of-day sign-off and memory consolidation routine.
How Mindkeeper Works
- The system initializes a standardized directory structure in the agent workspace through a bootstrap script.
- Layer 1 (MEMORY.md) acts as the primary context, loading tacit knowledge and preferences at the start of every session.
- Layer 2 captures raw, chronological data in daily markdown files to ensure no information is lost during interactions.
- Layer 3 organizes long-term data into Projects, Areas, Resources, and Archives for high-level retrieval.
- Supporting tools like QMD provide semantic search (BM25 and vector reranking) across all markdown layers.
- Automated consolidation jobs periodically review raw notes and promote distilled insights to the structured layers.
Mindkeeper Setup
To deploy Mindkeeper for your Openclaw Skills, execute the bootstrap script for your specific operating system to create the necessary directory structure.
Windows (PowerShell):
pwsh -File "<skill_dir>/scripts/setup.ps1" -Workspace "<workspace_path>"
Linux/Mac (Bash):
bash "<skill_dir>/scripts/setup.sh" "<workspace_path>"
After bootstrapping, configure the QMD search backend and the daily consolidation cron job as detailed in the provided reference documentation.
Mindkeeper Data Schema & Taxonomy
Mindkeeper utilizes a multi-layered data schema to maintain high-fidelity agent memory within Openclaw Skills:
| Component | Path | Function |
|---|---|---|
| Tacit Memory | MEMORY.md |
High-level patterns, user preferences, and distilled lessons |
| Daily Notes | memory/YYYY-MM-DD.md |
Raw capture of session events and quick insights |
| Projects | life/projects/ |
Active tasks with specific goals and deadlines |
| Areas | life/areas/ |
Long-term responsibilities and ongoing standards |
| Resources | life/resources/ |
Reference materials, how-to guides, and patterns |
| Archive | life/archive/ |
Completed projects and inactive knowledge assets |
Mindkeeper Advanced Features
- QMD Integration: Local semantic search using BM25, vectors, and reranking for precise retrieval.
- Daily Consolidation: Automated cron-based extraction of knowledge from session logs into the PARA structure.
- Transcript Verification: Enhanced voice-heavy workflows with inline transcript echo and local storage.
- Decision Records: Specialized ADR templates for tracking architectural and logic-based decisions over time.
- Multi-Agent Support: Standardized memory structure that can be shared or partitioned across different Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Black Forest Labs FLUX
A high-performance image generation skill that connects directly to Black Forest Labs' FLUX models via their official API.
Clawtales
A creative social platform connector that enables AI agents to publish serialized fiction and engage with a community of digital authors.
llm-wiki
An AI-driven personal knowledge base system that transforms fragmented information from various sources into a structured, interlinked markdown wiki.
会下指令 TaskFit
A local-first prompt optimizer that transforms vague user requests into high-quality, structured execution briefs for AI agents.
Agent Architect
An interactive architecture consultant that guides developers through designing and planning complex AI agent systems using proven patterns.
NoJargon Discovery Skill
A specialized discovery skill for NoJargon, the OpenClaw plugin that strips away technical jargon and buzzwords to deliver human-centric AI responses.
Featured*
