SuperClaude Framework v4: commands, agents, MCP servers

SuperClaude Framework is an open-source configuration layer for Claude Code, adding commands, agents, adaptive modes, and MCP integrations under a new /sc: namespace.

•
Claude
SuperClaude Framework v4: commands, agents, MCP servers

Overview

SuperClaude Framework is a meta-programming configuration project that positions Claude Code as a more structured development environment via behavioral instruction injection and component orchestration. The repository lists an MIT license, multi-language code (primarily Python), and distribution on both PyPI and npm. Project pages and packages: website at https://superclaude.netlify.app/, PyPI at https://pypi.org/project/SuperClaude/, and npm at https://www.npmjs.com/package/@bifrost_inc/superclaude.

Repository metadata on the project page indicates 14.2k stars, 1.3k forks, 120 watchers, 6 tags, and 31 contributors. Language breakdown shows Python 94.6%, JavaScript 3.8%, and Shell 1.6%. A star history chart is linked at https://www.star-history.com/#SuperClaude-Org/SuperClaude_Framework&Timeline, and the repo includes standard governance and logistics files (CODE_OF_CONDUCT.md, CONTRIBUTING.md, SECURITY.md, LICENSE).

Core feature set

The documentation outlines a framework centered on commands, agents, modes, and tool integrations:

  • 21 slash commands are exposed under a new /sc: namespace intended to avoid conflicts and organize the interaction surface across the development lifecycle.
  • 14 specialized agents are described, including roles such as “security engineer” and “frontend architect,” with notes on automatic coordination and domain-focused behavior based on context.
  • 5 adaptive behavioral modes are listed: Brainstorming, Orchestration, Token-Efficiency, Task Management, and Introspection. The Token-Efficiency mode cites “30–50% context savings.”
  • 6 MCP servers are integrated: Context7 (documentation context), Sequential (complex analysis), Magic (UI component generation), Playwright (browser testing), Morphllm (bulk transformations), and Serena (session persistence).

Version 4 emphasizes a smarter agent system, the /sc: command namespace, performance-oriented footprint reductions intended to allow larger projects and longer sessions, and a documentation overhaul with practical examples and clearer navigation.

Installation and upgrade notes

Installation is documented for multiple ecosystems:

  • Python via pipx: pipx install SuperClaude && pipx upgrade SuperClaude && SuperClaude install
  • Python via pip: pip install SuperClaude && pip upgrade SuperClaude && SuperClaude install
  • Node via npm: npm install -g @bifrost_inc/superclaude && superclaude install

The README flags PEP 668 scenarios and lists three approaches: use pipx, perform a user install (pip install --user SuperClaude), or use pip install --break-system-packages SuperClaude with caution. An upgrade note advises uninstalling V3 before installing V4, removing legacy SuperClaude V3 JSON files that may cause conflicts. The document also lists items preserved during upgrade (for example, custom slash commands outside commands/sc/, content in CLAUDE.md, and Claude Code settings JSON files).

Documentation and resources

The repository links a reorganized documentation set:

  • Getting started: Quick Start and Installation Guide
  • User guides: Commands Reference (21 commands), Agents Guide (14 agents), Behavioral Modes (5 modes), Flags Guide, MCP Servers (6 integrations), and Session Management (save/restore state)
  • Developer resources: Technical Architecture, Contributing Code, Testing & Debugging
  • Reference: Best Practices, Examples Cookbook, Troubleshooting

Docs are hosted in the repository under /Docs with direct links referenced in the README for each section.

Project details

The framework is MIT licensed (LICENSE in the root). The README highlights that the project welcomes contributions and outlines areas with “High,” “Medium,” and “Low” priority for community help, including documentation, MCP integrations, workflows, testing, and i18n. Support and sponsorship options are listed: Ko‑fi at https://ko-fi.com/superclaude, Patreon at https://patreon.com/superclaude, and GitHub Sponsors at https://github.com/sponsors/SuperClaude-Org. No pricing information is provided; the project is open source.

TL;DR

  • V4 focus: smarter agent system, /sc: command namespace, performance tuning, documentation overhaul
  • Scope: 21 commands, 14 agents, 5 modes, 6 MCP servers
  • Integrations: Context7, Sequential, Magic, Playwright, Morphllm, Serena
  • Platforms: installable via pipx, pip, or npm
  • License: MIT; repository lists 14.2k stars, 1.3k forks, 31 contributors
  • Docs: quick start, commands, agents, modes, MCP, and session management linked in README
  • Availability: open-source repository at https://github.com/SuperClaude-Org/SuperClaude_Framework and website at https://superclaude.netlify.app/

Continue the conversation on Slack

Did this article spark your interest? Join our community of experts and enthusiasts to dive deeper, ask questions, and share your ideas.

Join our community

Related Articles

Anthropic Working to Expand Claude Code to Browser with GitHub Integration

Anthropic Working to Expand Claude Code to Browser with GitHub Integration

Anthropic is testing a browser-based Claude Code with GitHub integration and sandboxed execution. The web tool blends AI copilots with cloud IDE workflows by connecting repos and running tasks in isolated environments.

1 shared tag
A Minimal AI Dev Workflow: Ghostty, Claude Code, and CLIs

A Minimal AI Dev Workflow: Ghostty, Claude Code, and CLIs

Developer Peter Steinberger outlines a lean AI coding setup centered on Ghostty, Claude Code, selective VS Code usage, and CLI-first tooling, with notes on planning, testing, and rate limits.

1 shared tag
Inside Claude Code’s Design: A Simplicity‑First Agent Playbook

Inside Claude Code’s Design: A Simplicity‑First Agent Playbook

A detailed community analysis outlines how Claude Code favors a single-loop architecture, small-model heavy lifting, and explicit prompts and tools over multi-agent or RAG complexity.

1 shared tag