Superset

Superset is an open-source agent orchestration harness and multi-agent workspace developed by the superset-sh team. It provides a structured environment for managing, orchestrating, and executing terminal-based AI coding agents across multiple projects and repositories. In day-to-day development, Superset provides a dedicated desktop and terminal workspace designed to run AI agents concurrently. It interfaces directly with agent protocols such as the Model Context Protocol (MCP) and supports modular Agent Skills specifications across Claude Code plugins and standard agent configurations (such as OpenAI YAML specs). Developers can execute one-off runs, run automated agent sweeps, or manage prompt configurations using built-in automation utilities. Rather than functioning solely as a standalone LLM wrapper, Superset differentiates itself as a unified orchestrator and agent runtime. It standardizes skill sharing across agent formats (e.g., symlinking skills between Claude and other agent runtimes), provides MCP wiring out of the box, and handles terminal process management to prevent UI locking during heavy multi-agent execution. Superset is fully open source under public repository licensing on GitHub, making it suitable for teams building customized internal agent workflows without proprietary platform constraints.

Tags: agent harness, ai agent, cli, code assistant, coding agent, developer tools

Visit Superset

Key features

  • Agent Orchestration Harness — Manage, run, and monitor multiple concurrent agent terminal sessions in a unified workspace. - MCP Integration — Connect agents natively to developer tooling, filesystems, and APIs using the Model Context Protocol. - Cross-Agent Skill Framework — Standardized skill configurations compatible with both Claude Code (.claude-plugin) and generic agent specifications (.agents/skills, OpenAI YAML). - Automated Workflow Execution — Built-in automation utilities (/superset-automation) to edit prompt bodies via standard editors and execute batch agent tasks. - Optimized Terminal Workspace — Desktop-oriented multi-terminal execution engine built to handle high-throughput agent logs and terminal sessions.

Use cases

Use Case: Cross-Agent Skill Management Maintain a single source of truth for agent skills across different CLI agents. Developers can define tools and instructions in .agents/skills and expose them seamlessly to Claude Code or custom OpenAI-compatible agent runtimes without rewriting configuration.

Use Case: Automated Batch Agent Sweeps Trigger automated repository maintenance or code-auditing sweeps across codebases. Using Superset automation commands and script harnesses, teams can execute structured tasks and collect feedback reports in a consistent format.

Use Case: Concurrent Multi-Terminal Development Run separate coding agents simultaneously across different microservices or branches in a managed multi-session workspace without terminal rendering locks or clashing state.

Who it is for

Superset is aimed at software engineers and engineering teams who actively run multiple AI coding agents, manage custom Agent Skills/MCP servers, and need an orchestration layer. It is not intended for users looking for a minimal, single-model CLI assistant without plugin or multi-session requirements.

Install guide

Installation Clone the repository and install dependencies locally (standard installation method for the Superset project source): bash git clone https://github.com/superset-sh/superset.git cd superset npm install To build and launch the application: bash npm run build npm start

First Run Configure your agent environment and link skills: 1. Ensure your AI agent environment variables (e.g., ANTHROPIC_API_KEY, OPENAI_API_KEY) are exported in your shell. 2. Configure your agent skills under .agents/skills/ or enable the included .claude-plugin for Claude Code. 3. Run the automation command to verify the setup: bash /superset-automation --help

Example Session Trigger an automation workflow or launch a managed agent task: bash # Launch an automated agent sweep with defined skill sets ./scripts/sweep.sh # Or configure a specific prompt body via your editor /superset-automation edit

Part of CLI Coding Agents

Categories: AI Coding Agents, Code Assistants, Developer Tools

Related CLI agents

  • Bwee — Bwee is a native macOS agent harness designed to organize terminal-based AI agents and render interactive user interfaces on the…
  • Okto Nexus — Okto Nexus is a local-first multi-agent coordination harness and orchestrator developed by OktoLabs. It acts as an execution bus designed…
  • Conductor — Conductor is a multi-agent orchestration harness developed by Melty Labs. It is designed to manage, coordinate, and execute parallel coding…
  • Defract — Defract is an agent orchestration harness and desktop workflow environment designed to coordinate multi-stage AI software engineering workflows. Built as…
  • Ona — Ona (formerly Gitpod's agentic platform, now part of OpenAI) is an enterprise-grade cloud agent orchestration platform that manages and runs…
  • Kiro — Kiro is an agentic engineering platform and CLI coding agent built and operated by AWS. It extends beyond standard code-completion…
  • AI Tools
  • Categories
  • Industries
  • CLI Coding Agents
  • MCP Servers
  • MCP Categories