Structured LLM task layer with retry, JSON parsing, and query-based routing

Different subs and CLI scripts for specific but repeatable LLM-based workflows.

talk to an MCP server, in either protocol era

MCP (Model Context Protocol) server framework for Raku

Ask-the-user tool for MCP::Server, built on server-side elicitation

Root-confined filesystem tools for MCP::Server

Allowlisted command-runner toolkit for MCP::Server, with capped output, streaming and background jobs

Raku client for OpenRouter's non-inference REST API

shai (SHell AI) - ask for LLM-powered shell commands.

WAT!? Does this code do? A handy LLM-powered command line utility to find out.