A robust memory and session management system that provides AI agents with a structured knowledge graph and resilient context recovery.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawvault
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 clawvault 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 ClawVault?
ClawVault is a comprehensive memory architecture designed specifically for Openclaw Skills to overcome the limitations of short-term context windows. It functions as a local, structured vault where agents can store and retrieve knowledge through a typed graph system. By utilizing markdown files and wiki-links, it builds a persistent web of decisions, lessons, and project states that persist across different sessions.
This skill serves as a central nervous system for your agents, allowing them to maintain continuity without losing focus. It includes sophisticated tools for context retrieval, session handoffs, and even automated repair of corrupted session transcripts. By integrating with these Openclaw Skills, developers can ensure their agents remain productive over long-term projects while maintaining a high degree of auditability and local data privacy.
ClawVault Use Cases
- Capturing and retrieving long-term project decisions and architectural lessons.
- Maintaining session continuity through structured handoffs, next steps, and blocker tracking.
- Repairing corrupted session transcripts caused by interrupted tool calls or API errors.
- Building a searchable knowledge graph of people, projects, and interconnected entities.
- Preventing context death by periodically checkpointing the agent's current state during heavy workflows.
How ClawVault Works
- The user initializes a local vault which creates a standardized directory structure and document templates.
- At the start of a session, the wake command is invoked to recap recent handoffs and bootstrap the agent's memory.
- During active development, the agent uses checkpoint and capture commands to record progress and atomic pieces of information.
- Specific memory types like decisions, lessons, or relationships are categorized into dedicated folders with associated metadata.
- A knowledge graph is incrementally built from wiki-links and frontmatter, enabling graph-aware context retrieval.
- Upon session completion, a sleep command generates a transition summary, ensuring the next agent instance can resume work immediately.
ClawVault Setup
To integrate this powerful memory system into your Openclaw Skills environment, follow the canonical installation process:
npm install -g clawvault
npm install -g github:tobi/qmd
openclaw hooks install clawvault
openclaw hooks enable clawvault
Initialize your vault and configure the environment:
# Initialize the vault directory
clawvault init ~/my-vault
# Set the path environment variable
export CLAWVAULT_PATH=~/my-vault
ClawVault Data Schema & Taxonomy
ClawVault uses a highly organized directory structure to manage agent data, making it easy for both humans and agents to navigate knowledge:
| Path | Content Type | Purpose |
|---|---|---|
/decisions |
Markdown | Logs key choices, reasoning, and alternatives considered. |
/lessons |
Markdown | Captures insights, patterns, and anti-patterns for future reference. |
/people |
Markdown | Tracks contacts, roles, and relationship history. |
/projects |
Markdown | Manages active work tracking and specific project context. |
/handoffs |
Markdown | Stores session summaries, blockers, and next-step instructions. |
/inbox |
Markdown | Acts as a landing zone for quick, unprocessed captures. |
.clawvault/ |
JSON/Hidden | Stores the internal graph index and session death flags. |
ClawVault Advanced Features
- Session Transcript Repair: Automatically fixes orphaned tool blocks and broken parent chains in Openclaw session histories.
- Graph-Aware Context: Specialized retrieval profiles (planning, incident, handoff) that pull information based on entity relationships.
- LLM-Powered Observation: Uses optional compression to turn verbose session transcripts into concise, actionable observations.
- Wiki-Link Automation: Automatically links entity mentions across the vault to build a dense, navigable knowledge graph.
- Comprehensive Diagnostics: Built-in doctor and compatibility commands to ensure the hook and runtime are perfectly aligned.
SKILL.md
Loading
Related Openclaw Skills
Bambu Lab 3D Printer Control
A local-only CLI tool enabling AI agents to monitor and control Bambu Lab 3D printers via MQTT and FTP protocols.
Agent Memory Templates
A production-tested framework for managing AI agent personality, memory checkpoints, and session persistence.
Agent Autonomy Primitives
A standardized framework of five composable primitives designed to turn AI agents into persistent, self-directing autonomous workers.
Web Automation & Browser Workflows
A comprehensive skill for automating complex web interactions, form filling, and data extraction via a cloud-based REST API.
CloverCLI Skill
A comprehensive command-line interface for managing Clover POS operations, from inventory tracking to advanced financial reporting.
LinkedIn via Pipedream
A powerful automation skill for managing LinkedIn profiles and organization pages through Pipedream OAuth integration.
METADATA
Featured*
