Open Multi-Agent Kit (OMK) is an open-source terminal framework and multi-agent orchestration harness developed by dmae97. It is built to coordinate specialized coding subagents, route tasks across language models, and safely execute code generation and refactoring workflows directly from the terminal. OMK functions through a modular architecture composed of core packages including omk-agent-core, omk-ai, and an interactive terminal user interface (omk-tui). In daily workflows, tasks are decomposed into dependency graphs (TaskGraphPlan) that delegate subtasks to individual agents. The harness maintains context via local repository wiki metadata (such as AGENTS.md and CLAUDE.md) and prompt attachment stores, while enforcing strict execution security through a centralized command safety gate with recursive substitution checks. What sets OMK apart from standard single-loop CLI agents is its focus on subagent graph execution, runtime provenance tracking, dynamic reasoning routing, and integration with the Model Context Protocol (MCP). It features automated tool recovery and retry loops to handle complex developer workflows without failing on isolated execution errors. The project is fully open source under active development on GitHub, distributed as a modular npm workspace with support for major model APIs and OAuth authentication providers.
Subagent graph orchestration — Decomposes complex coding tasks into structured execution graphs (TaskGraphPlan) with task delegation and retry recovery. - Interactive terminal UI — Terminal workspace (omk-tui) featuring prompt attachment strips, session selection, and scrollback rendering. - Command safety gate — Centralized execution gate that parses recursive shell command substitutions to prevent unsafe tool executions. - Reasoning router — Dynamic model routing engine that assigns subtasks to optimal LLMs based on task complexity and model capability weights. - MCP integration — Integrates Model Context Protocol tools to connect external services and developer tooling into the agent runtime. - Repository wiki ingestion — Ingests project-level architectural context and rules directly from files like AGENTS.md and CLAUDE.md.
Use cases
Use Case: Complex Multi-Agent Refactoring Developers can initiate repository-wide refactors where OMK decomposes the workflow into task graphs. Dedicated subagents handle dependency analysis, code modifications, and test suite execution in coordinated stages.
Use Case: Terminal-Based Interactive Development Engineers can use the omk-tui interface for daily terminal pairing, attaching targeted file snippets into prompt strips and reviewing agent diffs interactively.
Use Case: Extensible MCP Tool Execution Teams can integrate custom MCP servers into OMK to allow coding agents to query external databases, issue tracking systems, or local build environments during automated code generation.
Who it is for
OMK is suited for software engineers and systems developers who want a modular, terminal-first multi-agent harness with strong command safety controls and graph-based task planning. It is not intended for non-technical users or developers looking for a lightweight single-prompt chat plugin.
Install guide
Installation Install the OMK suite globally using npm: bash npm install -g open-multi-agent-kit omk-agent-core omk-ai omk-tui
First Run Set your model API keys or configure OAuth credentials: bash export OPENAI_API_KEY="your-api-key" export XAI_API_KEY="your-xai-key" # Initialize OMK in your target repository omk init
Example Session Launch the terminal interface or run a task directly through the CLI: bash # Launch interactive TUI omk-tui # Or execute a headless graph task omk "Inspect src/auth/jwt.ts, identify potential expired token bypasses, and add test cases to test/jwt.spec.ts"
SWE-agent — SWE-agent is an open-source autonomous software engineering agent harness built and maintained by researchers from Princeton University and Stanford University.…
Agent Runbook — Agent Runbook is an open-source contract-based multi-agent skill framework developed by KnoxOps. It acts as an orchestrator and compiler that…
Weaver — Weaver is an open-source agent harness and orchestration framework maintained by sean35mm on GitHub. Built on top of the Bun…
Context Bridge — Context Bridge is an open-source CLI agent harness and context orchestration utility developed by serdardb. It is built to facilitate…
Hivelore — Hivelore is an open-source agent harness and contextual memory orchestrator created by Doucs91. It provides persistent project memory, decision tracking,…
Agentic Engineering Framework — Agentic Engineering Framework is an open-source orchestration harness and governance layer developed by Dimitri Geelen. It is designed to constrain,…