Agentswarm — DevOps skill for Claude Code

agentswarm

[](https://www.npmjs.com/package/@robzilla1738/agentswarm) [](LICENSE) [](package.json) [](https://buymeacoffee.com/robcourson)

A local agent-swarm orchestrator with a terminal dashboard and a localhost web UI. Works with DeepSeek, OpenAI, Anthropic, xAI, MiniMax, OpenRouter, Ollama, LM Studio, or any OpenAI-compatible endpoint.

You give it a mission. A conductor model breaks the mission into tasks and hands them to worker agents that run in parallel, share findings on a blackboard, and get checked by an adversarial verifier. The run ends with a synthesized report plus whatever files the agents produced. Everything runs on your machine with your own API key, or fully offline against a local model.

            ┌─────────────┐
            │  Conductor  │  decomposes the mission, schedules waves,
            └──────┬──────┘  reacts to results, steers toward the goal
       ┌───────────┼───────────┐
   ┌───▼───┐   ┌───▼───┐   ┌───▼───┐    parallel worker agents
   │  T1   │   │  T2   │   │  T3   │    (shell · files · web · notes)
   └───┬───┘   └───┬───┘   └───┬───┘
       └─────┬─────┘           │         dependencies + shared blackboard
        ┌────▼─────┐      ┌────▼────┐
        │  T4 dep  │◀─────│ verify  │   adversarial verification
        └────┬─────┘      └─────────┘
        ┌────▼─────┐
        │Synthesize│  → final report (.md + .html) + artifacts
        └──────────┘

What it does

• Independent tasks run at the same time, up to a parallelism cap you