config-knowledge

name	config-knowledge
description	Knowledge Manager — Vector-based semantic search, KB indexing, and cache management
metadata	{"version":"1.0.0","author":"Network AI Team","type":"agent","category":"knowledge-management","intent":"semantic_search_knowledge_base"}
tools	["search_knowledge","index_knowledge_files","get_knowledge_status"]
system	$ref:./prompts/system.md

Overview

The Knowledge Manager handles OLAV's semantic search and knowledge base operations.

Use Cases

1. Semantic Search: Find relevant information using natural language
2. KB Indexing: Index new documents (Markdown, PDF) into the vector store
3. Cache Management: Manage the semantic cache for fast responses
4. Index Health: Monitor KB index status and statistics

Configuration

• Knowledge directory: .olav/knowledge/
• Vector DB: .olav/databases/main.duckdb
• Default search limit: 5 results
• Default chunk size: 1024 characters

Best Practices

1. Index new documents after adding to .olav/knowledge/
2. Use incremental indexing to avoid duplicates
3. Monitor index health regularly
4. Use semantic search before performing deep analysis