Minimalist Knowledge Base MCP

# Instructions for LLMs Working with Personal Knowledge Base MCP This document provides guidance for Large Language Models interacting with the Personal Knowledge Base MCP server. ## Available Tools 1. createArticle(content: string, keywords?: string[]) - Purpose: Create a new article - Returns: { id: string, filePath: string } - Example usage: Create articles with structured content and relevant keywords 2. getArticle(id: string) - Purpose: Retrieve article content - Returns: { content: string, filePath: string } - Use when: Need to read or reference existing articles 3. updateArticle(id: string, content: string, keywords?: string[]) - Purpose: Modify existing articles - Returns: { success: boolean } - Best practice: Preserve existing structure while updating content 4. deleteArticle(id: string) - Purpose: Remove articles - Returns: { success: boolean } - Use carefully: Operation cannot be undone 5. searchArticles(query: string, timeRange?: {start?: string, end?: string}, timeField?: "created"|"modified") - Purpose: Full-text search across articles - Returns: Array of matching articles with relevance scores - Best practice: Use specific search terms for better results 6. findLinkedArticles(id: string) - Purpose: Find articles sharing keywords - Returns: Array of related articles - Use for: Discovering content connections 7. getArticlesByTimeRange(startTime?: string, endTime?: string, type: "created"|"modified") - Purpose: Time-based article retrieval - Returns: Array of articles within the time range - Time format: ISO 8601 8. listArticles() - Purpose: Get overview of all articles - Returns: Array of article metadata - Use for: Understanding knowledge base structure 9. getArticleStats(timeRange?: {start?: string, end?: string}) - Purpose: Analyze knowledge base metrics - Returns: Statistics about articles and keywords - Use for: Understanding content patterns ## Best Practices 1. Content Structure - Use Markdown formatting - Include a clear H1 title - Organize content with appropriate headings - Use lists and code blocks when relevant 2. Keywords - Choose specific, relevant keywords - Use consistent terminology - Avoid overly generic terms - Consider related concepts 3. Search Optimization - Use specific search terms - Consider temporal context when relevant - Combine with keyword search for better results 4. Error Handling - Check return values for success - Handle missing articles gracefully - Provide clear error messages to users ## Content Guidelines 1. Article Format ```markdown # Title ## Overview Brief introduction ## Main Content Detailed information ## Keywords - keyword1 - keyword2 ``` 2. Metadata Handling - Preserve existing metadata when updating - Update keywords when content focus changes - Maintain consistent formatting 3. Cross-Referencing - Use keywords for implicit linking - Consider content relationships - Maintain bidirectional references when appropriate ## System Interaction 1. Initialization - Server must be running - Knowledge base must be specified - Tools are available after connection 2. Performance Considerations - Batch related operations - Use search before retrieving full content - Consider pagination for large result sets 3. Data Consistency - Verify operations complete successfully - Handle concurrent modifications - Maintain content structure Remember: You are working with a file-based system where Markdown files are the source of truth. Always ensure content is properly formatted and structured for long-term maintainability.