Index local Markdown files and retrieve semantic search results with contextual information.
npx skills add https://github.com/dlants/pkb --skill pkb
Please help me install this Agent Skill. Skill: pkb Source: https://github.com/dlants/pkb/tree/main Command: npx skills add https://github.com/dlants/pkb --skill pkb
PKB provides a local semantic search for Markdown documents, reducing time spent finding information across notes and enabling knowledge reuse.
High-intent search queries and answers about installing and using this skill.
Local semantic search over Markdown notes requires chunking file content, generating embeddings via an embedding model, and storing them in a SQLite database with a vector extension to enable contextual retrieval.
Yes, you can index individual Markdown files or entire directories. The CLI supports tracking specific file paths and syncing all tracked sources to build a searchable local knowledge base.
Running a local vector database for contextual retrieval requires SQLite with the sqlite-vec extension for storage, better-sqlite3 for database access, and an embedding model to generate vector representations.
Contextual retrieval works by chunking Markdown content into smaller segments, embedding those chunks into a vector database, and querying the embeddings to find relevant results with their surrounding contextual information.
Yes, the sqlite-vec extension is required. It provides the vector database functionality within SQLite needed to store and query the embeddings generated from your Markdown documents.
The best way is to index your Markdown files into a local SQLite vector database, allowing AI agents to query the embedded chunks and retrieve relevant contextual information from your notes.
mizchiConfigure Node 24's built-in sqlite with the sqlite-vec extension.
sixscripts-aiOptimize retrieval-augmented generation pipelines for LLM-powered tasks.
aurainfosecInvoke AWS Bedrock foundation models for text generation and embeddings.
QredenceIngest markdown files into ChromaDB for semantic procedural memory retrieval.
