Claw Local Knowledge

A specialized skill enabling AI Agents to manage a local document knowledge base with semantic search and retrieval capabilities.

Overview

This skill provides two core functionalities:

Function	Description
Ingest Knowledge	Convert user-uploaded documents (docx, pdf, xlsx, pptx) into markdown and store them in the knowledge base
Retrieve Knowledge	Search and retrieve relevant documents from the knowledge base to assist decision-making

Architecture

.openclaw/workspace/
├── memory/
│   ├── knowledge_base/              # Converted markdown documents
│   └── knowledge_base_index.json     # Knowledge base index (JSON array)
└── temp_file/                        # Staging area for uploaded files

File Flow

1. User uploads documents → .openclaw/workspace/temp_file/
2. Agent converts to markdown → .openclaw/workspace/memory/knowledge_base/
3. Agent updates index → .openclaw/workspace/memory/knowledge_base_index.json

Initial Setup

Before first use, initialize the required directory structure:

Linux / macOS

# Create knowledge base directory
mkdir -p .openclaw/workspace/memory/knowledge_base

# Create temporary file directory
mkdir -p .openclaw/workspace/temp_file

# Create empty index file
echo '[]' > .openclaw/workspace/memory/knowledge_base_index.json

Windows PowerShell

# Create directories
New-Item -ItemType Directory -Path ".openclaw/workspace/memory/knowledge_base" -Force
New-Item -ItemType Directory -Path ".openclaw/workspace/temp_file" -Force

# Create empty index file
Set-Content -Path ".openclaw/workspace/memory/knowledge_base_index.json" -Value "[]"

Configure SOUL.md

If the SOUL.md file exists in the OpenClaw workspace, inject the following instructions in the original SOUL.md style to enable proactive knowledge retrieval:

## Local Knowledge (claw-local-knowledge skill)
Please attempt to load the **claw-local-knowledge skill** when encountering uncertain or specialized knowledge, and retrieve relevant information according to the instructions.

**Retrieval Process:**
1. Read `memory/knowledge_base_index.json` to obtain the file list.
2. Match relevant files based on keywords.
3. Read the corresponding Markdown files from `memory/knowledge_base/`.
4. Integrate the information and reply to the user.

**Note:**
The knowledge base path is `workspace/memory/knowledge_base/`, and the index file is `workspace/memory/knowledge_base_index.json`.

This configuration ensures that the Agent can proactively query the local knowledge base when facing knowledge blind spots.

Index File Format

The knowledge_base_index.json is a JSON array where each entry represents a document:

{
  "name": "document_name.md",
  "description": "Brief summary of the document's content"
}

Supported Input Formats

Format	Type	Extension
Word Document	Microsoft Word	.docx
PDF Document	Portable Document Format	.pdf
Excel Spreadsheet	Microsoft Excel	.xlsx
PowerPoint Presentation	Microsoft PowerPoint	.pptx

Usage Scenarios

• User uploads documents and requests integration into the knowledge base
• Agent requires domain knowledge to assist with user queries
• User inquires about topics covered by existing knowledge base documents
• Agent encounters uncertainty and needs to verify against stored documentation

Workflow Reference

For detailed workflow instructions, refer to:

• references/add_knowledge.md — Document ingestion process: conversion, cleaning, indexing
• references/retrieval_knowledge.md — Knowledge retrieval process: index lookup, content retrieval, integration