omk
Active·★ 148·MIT·Updated 2026-10-22
OMK, or Open Multi-Agent Kit, provides a provider-neutral coding-agent CLI and a multi-agent control plane. It aims to scope work, route appropriate agents, and verify every release for reliable engineering outcomes.
#Multi-Agent#Coding Agent#Orchestration#CLI#LLM Control Plane#Task Automation#Code Verification#Developer Tooling
01
Features
01Multi-agent execution with bounded DAGs and owned paths to prevent conflicts and manage resources.
02Verifiable outcomes using acceptance predicates backed by fresh command evidence, ensuring task completion is genuinely confirmed.
03Provider-neutral LLM integration allowing flexible routing to various models like Codex, Claude, Grok, and local providers.
04Durable goals and session recovery from checkpointed states, enabling resilience and replayability for interrupted workflows.
05OS-level sandboxing for local bash operations, restricting writes to owned paths and disabling network access by default for enhanced security.
02
Compatibility
Node.js
Runtime
Verified via docs
Linux
OS Sandboxing
Verified via docs
macOS
OS Sandboxing
Verified via docs
LLM Providers
Integrations
Verified via docs
03
Quick start
1
$ npm install -g open-multi-agent-kit --ignore-scripts
04
Use cases
↳Orchestrating complex coding tasks and multi-step software development projects using collaborative AI agents.
↳Ensuring code quality and correctness by automatically verifying agent-generated code against defined build, type, test, and audit gates.
↳Building resilient AI-driven workflows that can recover from failures and continue tasks from where they left off.
↳Managing costs and performance by dynamically routing tasks to different LLM providers based on efficiency or specific capabilities.
↳Securely executing untrusted agent code by running commands within isolated, sandboxed environments to mitigate security risks.
05
Alternatives
claw-orchestrator★ 596
Run Claude Code, Codex, Antigravity, Grok Build and OpenCode behind one runtime, and a run only counts as done when checks it didn't write pass. Cross-engine fan-out, councils, session handoff, spend caps, MCP server, ACP agent.
ask-user-questions-mcp★ 147
Better 'AskUserQuestion' - A lightweight MCP server/OpenCode plugin/Agent Skills + CLI tool that allows your LLMs ask questions to you. Be the human in the human-in-the-loop!
agentao★ 309
Local-first, governed AI agent runtime for Python — embed it in your app, or run it as a CLI or ACP server. Permissions, MCP, memory and audit replay built in.
oto-dock★ 193
The self-hosted platform for your company's AI agents. Your team builds them on Claude Code, Codex or local models, and they work in chats, on schedules, on the phone and on your own machines.
nac★ 284
Give AI agents ambitious work without losing the plot. nac is an open-source harness for long-running tasks, using a central orchestrator, threads, and structured episodes to stay aligned with your intent.
Comments
Log in to leave a comment
No comments yet. Be the first!