Ai Devkit

AI DevKit is an open source command line tool that acts as a control plane for developers running several AI coding agents at once. Rather than replacing Claude Code, Codex, Cursor, Gemini CLI, opencode or Pi, it sits above them and gives them one shared configuration, one console, local memory, a way to pass work between agents, and verification before a task is marked done. It is built for developers who already operate more than one coding agent and find the setup scattered across terminal tabs and duplicated rule files. A single project file reconciles agent integrations from one source of truth. The agent console lists running agents and local sessions. The agent send command routes prompts, logs, test output or review requests to a specific agent or a saved group, including piped input from a test run. A local SQLite memory stores conventions, decisions and reusable fixes that agents search only when they need them, which keeps prompts lean. Workflow skills and verification gates require fresh build or test output before an agent can claim completion, and a library of reusable skills covers planning, debugging, documentation and review. The tool is free. It is published on GitHub and distributed through npm, with a setup step run once per machine to connect detected agents and an init step run in each project to write reviewable, committable files. Everything runs locally, so memory and configuration stay on the developer's machine.