AI DevKit
A CLI to run several coding agents: one config, a session console, local memory and a set of engineering skills
Install
npx ai-devkit@latest initThe wizard asks which agents to configure.
This is third-party code. Review the repository files before installing.
What it does
The init command creates .ai-devkit.json and configures chosen agents from it: skills, MCP servers and a docs/ai folder with phases from requirements to testing. agent list and agent console show running sessions across agents, and agent send passes them a prompt or command output. Memory is stored locally in SQLite and exposed via CLI and MCP. Bundled skills include dev-lifecycle, verify, tdd, review, debugging, security and docs, plus driving an agent from Telegram or Slack.
Who it is for. For developers running several agents at once who want order in their setup and workflow.
Good fit when
- You need the same setup for Claude Code, Codex and Cursor in one repo
- You want a feature to go through requirements, design, plan, implementation and review
- You need to pass logs and tasks to running agents without copy-paste
Not a fit when
- You need a quick fix without process or docs
- Agent housekeeping files are not allowed in the repo
Example request
Run the new report export feature through dev-lifecycle, starting with requirementsLimitations
Requires Node.js. The dev-lifecycle process adds phase documents to the project and needs discipline. Some features, such as agent groups, are not supported by every provider. Telegram control needs a bot and Telegram access.
How to disable. Delete .ai-devkit.json, the installed skills in agent folders and the MCP entries in their settings.
Security check
- Sends instructions to running agents, including remotely from Telegram and Slack
- Changes agent and MCP settings in the project
README in short
The README presents AI DevKit as a control layer for many agents and shows what lands in the repo after init. It lists console and messaging commands, memory usage and the skill set led by dev-lifecycle. Building from source and docs on the website are covered. Apache-2.0 licensed.
SKILL.md
--- name: dev-lifecycle description: AI DevKit · Orchestrator for structured SDLC phase skills. Use when the user wants to run the full lifecycle or choose the next phase across requirements, design, planning, implementation, testing, and review. --- # Dev Lifecycle Coordinate the phase-specific AI DevKit skills instead of running phase details directly. Required phase skills: - `dev-worktree` for feature workspace setup and resume. - `dev-requirements` for phases 1-2: new requirement and requirements review. - `dev-design` for phase 3: design review. - `dev-planning` for phases 4 and 6: initial task planning and updates after implementation tasks. - `dev-implementation` for phases 5 and 7: execute plan and check implementation. - `dev-testing` for phase 8: write tests and verify coverage. - `dev-review` for phase 9: final code review. Supporting skills: - `memory` for reusable project knowledge during clarification. - `tdd` for implementation tasks. - `verify` before completing implementation, implementation checks, testing claims, and review readiness. - `task` for optional progress tracing when the task command is usable.
FAQ
Where is memory stored?
Locally in SQLite; agents search it on demand via MCP or CLI instead of keeping it all in context.
Can I add third-party skills?
Yes, ai-devkit skill add <registry> <skill> pulls skills from external registries.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Anthropic's coding agent for the terminal, IDE and GitHub: understands your codebase, runs tasks and handles git