Experimental skills for Claude Code Superpowers, this plugin extends Claude Code's capabilities with new techniques and tools that are still under active development and refinement. It targets users of Claude Code seeking to enhance its functionality with advanced, experimental features, offering benefits like semantic code analysis, on-demand tool invocation, interactive CLI automation, and a headless Windows VM environment.

How It Works

The finding-duplicate-functions skill employs a two-phase approach: classical function extraction followed by LLM-powered intent clustering to detect semantically similar code, not just syntactic duplicates. The mcp-cli skill allows on-demand querying and invocation of MCP servers without permanent integration. using-tmux-for-interactive-commands automates traditionally manual workflows by controlling interactive CLI tools via programmatic keystroke injection into detached tmux sessions. Lastly, windows-vm provisions and manages headless Windows 11 VMs within Docker, leveraging KVM acceleration and providing SSH access for cross-platform development and testing without requiring RDP.

Quick Start & Requirements

Install the plugin using claude-code plugin install https://github.com/obra/superpowers-lab or by adding the URL to your claude.json plugins configuration. tmux must be installed on your system. Skills are tested on Linux/macOS.

Highlighted Details

• finding-duplicate-functions identifies code with the same intent but different implementations, complementing syntactic duplicate detectors.
• using-tmux-for-interactive-commands enables automation for interactive tools like vim, git rebase -i, and REPLs.
• windows-vm provides a headless Windows 11 Docker VM with KVM acceleration and SSH, pre-configured with OpenSSH Server, Node.js, and Claude Code.

Maintenance & Community

The project encourages contributions via issues or pull requests. Related projects include superpowers (core skills library) and superpowers-chrome (browser automation skills).

Licensing & Compatibility

The project is released under the MIT License.

Limitations & Caveats

Skills are experimental, functional but under active refinement, and may evolve based on usage and feedback. Testing has primarily been conducted on Linux and macOS environments.