rbw-claude-code

Universal AI-powered development tools: 29 agents, 18 commands, 16 skills for code review, research, workflow automation, and instruction hygiene

16 skills · 18 commands · 1 MCP server · hooks

Python-specific tools: 7 agents, 2 commands, 1 skill for Python code review, testing, and type checking

2 skills · 2 commands

Advanced research plugin combining deep web research with GDELT-powered geopolitical analysis. Includes general research agents, specialized geopolitical analysts, and MCP server for structured GDELT queries. Diffusion-based deep research using parallel subagents, iterative gap detection, and comprehensive synthesis. Implements the ThinkDepth research loop natively in Claude Code.

1 skill · 1 MCP server

Block messy patterns (python -c, heredocs), nudge toward clean alternatives

hooks

Validate git commit messages follow conventional format

hooks

Block bare python/pip/pytest commands, enforce uv usage

hooks

Block Gemini 2.x models, enforce Gemini 3 models only

hooks

Allow only safe read-only gh api commands

hooks

Block destructive git commands to prevent data loss

hooks

Quickshell (Qt6/QML) desktop shell development for Hyprland: API reference, patterns, visual effects, gotchas, and feature matrix across reference shells

1 skill

Prevent reading .env files to protect secrets

hooks

Auto-format Python files with ruff after edits

hooks

Run type checking after Python file edits

hooks

Block destructive file operations, env file reading, and supply chain attacks

hooks

Remind to add tests when creating new Python files

hooks