AIfred Toolkit

AIfred Toolkit (aifred-tk) is a Python toolkit that exposes AI agent tools through a microkernel plugin system, available via both a CLI and an MCP (Model Context Protocol) server.

What is aifred-tk?

aifred-tk provides a uniform interface for composable AI tools:

• Plugin system — tools are registered via Python entry points and discovered at runtime
• CLI layer — every registered tool becomes a subcommand with typed options
• MCP layer — the same tools are exposed as an MCP server for agent frameworks

Sections

• Installation — get aifred-tk running
• Configuration — layered settings and per-tool options
• CLI Reference — command-line usage
• MCP Server — stdio and HTTP transports
• Elicitation — runtime user-choice API for plugins and LLM agents
• Ethics and Responsibility — AI governance and model attribution
• daimyo Integration — expose aifred-tk tools to daimyo rule templates
• Plugin System — how the plugin system works
• Creating Plugins — write your own tools
• Best Practices — architectural patterns for robust tools
• Conventional Commits — AI-assisted commit message generation
• File Analysis — LLM-powered file content analysis
• Prompt Injection Detector — detect prompt injection attempts in text
• Translation — LLM-powered text and file translation