By LieBieS
Newt is a quirky AI development assistant providing architecture analysis, automated code reviews, security auditing, performance analysis, code quality insights, PR review automation, and structured brainstorming.
Audits UI code and markup for WCAG compliance and practical usability.
Produces a deterministic ADR draft suitable for pasting into your repo's ADR directory.
Performs deterministic architecture validation against Clean, Hexagonal, Layered, Modular Monolith, and Microservices blueprints. Ideal for early warning during large refactors.
Runs a deterministic, high-signal brainstorming session that generates options and converges to a small number of testable candidates. Produces decision artifacts (Idea Cards, ADR Draft, Experiment Brief) and logs the session.
Reviews an existing component (or drafts a new spec) with a focus on web UI best practices.
Evaluates repository topology and enforces architectural fitness across popular enterprise styles.
Primary facilitator for Newt's Ideation OS; runs structured ideation, clustering, scoring, and produces decision artifacts.
Identifies explicit/hidden constraints and assumptions; proposes constraint relaxation options to expand the solution space safely.
Scores and selects the best ideas using a deterministic rubric; outputs top candidates and clear trade-offs.
Cross-domain inspiration agent that proposes analogies, patterns to borrow, and modern practices to adapt.
Analyze headings, landmarks, and primary CTA clarity to improve visual hierarchy and scannability in web UIs.
Validate that the repository is in a clean, buildable state before CI/CD pipeline execution.
Analyzes staged changes to suggest optimal commit boundaries and logical groupings.
Check UI component consistency for spacing, color tokens, typography scale, and variant duplication.
Extracts explicit and implicit constraints from configuration changes and architecture findings to keep an up-to-date constraints register.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
AI-Powered Development Assistant for architecture, security, performance, and quality automation
"Game over, man! Game over!" - Don't let bugs be the Xenomorphs of your codebase. Newt's got your back, just like LV-426's finest.
Newt is a comprehensive AI development assistant plugin that transforms how you review, plan, and improve code. Built for Claude Code, Windsurf, and Cursor, it provides intelligent automation across your entire development lifecycle.
Named after the plucky survivor from LV-426, Newt is your reliable companion in the hostile environment of software development - because sometimes the only way to survive is with a smart AI watching your six.
| 🔍 Code Review Automation | 🚀 PR Review Intelligence | 💡 Structured Brainstorming | ⚡ Continuous Quality |
|---|---|---|---|
| Architecture analysis | Commit planning | Ideation sessions | Real-time suggestions |
| Security audits | PR splitting | Decision artifacts | Automated skills |
| Performance insights | Review-ready summaries | ADR generation | Quality monitoring |
| Xenomorph detection | Dropship deployment | Hudson's optimism | LV-627 survival |
| Bug hunting in the vents | Smooth CI/CD pipelines | Mostly positive vibes | Code that survives production |
| 📖 Section | 🎯 Purpose | 🔗 Quick Link |
|---|---|---|
| Quick Start | Get up and running in minutes | ⚡ Quick Start |
| Available Commands | All Newt commands and usage | 📋 Available Commands |
| Agents & Skills | AI agents and their capabilities | 🤖 Agents • 🛠️ Skills |
| Architecture | System design and flow | 🏗️ Architecture |
| Installation | Detailed setup instructions | 📦 Installation |
| Configuration | Customize Newt for your needs | ⚙️ Configuration |
| LSP Server | Language Server Protocol integration | 🔌 LSP Server |
| Examples | Real-world usage scenarios | 💡 Examples |
| Troubleshooting | Common issues and solutions | 🔧 Troubleshooting |
| Learning Hub | Understand AI concepts (LLM, Agents, MCP, RAG) | 📚 Learning Hub |
| Hooks System | Automated workflow integration | 🪝 Hooks System |
| Templates | Project templates and scaffolding | 📋 Templates |
| Contributing | How to contribute to Newt | 🤝 Contributing |
| Support | Get help and community resources | 🆘 Support |
# Clone or download the plugin
cd /path/to/your/workspace
# Install in Claude Code/Windsurf/Cursor
/plugin marketplace add ./newt
/plugin install newt
# Verify installation
/review --help
# Run a quick code review
/review --path src/auth --depth quick
# Check project health
/project-health
# Review your staged changes before commit
/pr-review --staged
Edit config/default.yml to customize thresholds, policies, and integrations.
📖 See docs/installation-guide.md for detailed setup instructions.
Newt ships with a native MCP server under mcp/server.mjs for seamless integration with MCP-capable clients like Claude Desktop.
npm install
npm run mcp:server
{
"mcpServers": {
"newt": {
"command": "node",
"args": ["mcp/server.mjs"],
"cwd": "/absolute/path/to/newt"
}
}
}
| Resource | Description |
|---|---|
config://default.yml | Main configuration file |
config://schema.json | Configuration schema |
logs://reviews/latest | Latest review logs |
logs://brainstorm/latest | Latest brainstorm sessions |
agents://list | Available agents |
skills://list | Available skills |
Newt MCP tools return deterministic runbooks for execution in your agentic IDE:
npx claudepluginhub liebies/newt --plugin newtMulti-perspective code analysis covering architecture, security, and best practices
Live codebase visualization and structural quality gate — 14 health dimensions graded A-F, dependency analysis, and architecture governance via MCP
Personal Claude Code + Codex dev stack: security hooks, AI-first code conventions, /security-review, /repo-map, /stack-check, portable statusline. Designed to complement other skills-based plugins, not replace them.
Automated code review with severity levels and actionable feedback
Codebase exploration, refactoring, and quality analysis
Reflective code review toolkit with structured analysis, blast radius analysis, domain expertise, and intelligent review orchestration