Jennie 🤖⚡

The Autonomous AI Code Reviewer for Indie Devs & Teams

Bugs, leaked secrets, missing tests, and architectural risks — caught before they merge.

💡 What is Jennie?

Jennie is an autonomous, agentic AI code review tool built for modern engineering workflows. Unlike traditional static linters (ESLint, SonarQube) that only check syntax rules, Jennie performs full deep read-throughs like a senior engineer.

It traces AST call trees, follows cross-file dependencies, respects custom repository guidelines (.jennie/rules.md), connects to live external systems via the Model Context Protocol (MCP), and posts human-quality inline pull request feedback directly on GitHub.

Jennie can be run locally in your terminal before committing, continuously as a GitHub Action, or as an ambient background QA daemon.

🔥 Key Features

• 🔒 Whole-Codebase Security Audit: Run npx @tarunagnihotri534/jennie audit to scan your entire repository for sensitive points, leaked API keys (sk-..., ghp_..., AIza...), unprotected API route handlers, and unsafe code injections.
• 🧠 Deep Codebase Context: Explores far beyond git diffs. Traces imports, inspects modified call trees, and verifies interface contracts across files.
• 🔌 Model Context Protocol (MCP): Native MCP client allows connecting live databases, Sentry observability, and Playwright browser suites into the review loop.
• ⚡ Provider Agnostic: Native support for Anthropic (Claude 3.7 Sonnet / 3.5 Haiku), OpenAI (gpt-4o / o3-mini), OpenRouter, and Cloudflare Workers AI.
• 📜 Custom Repository Rules: Place a .jennie/rules.md file in your repository to enforce team-specific architecture, styling, and security requirements.
• 💬 On-Demand PR Comment Triggers: Trigger reviews directly inside GitHub PR comments using /jennie review.
• 🛡️ Security & Privacy First: Zero code storage. Your code diffs are processed directly by your chosen LLM provider and never used for model training.
• 🌐 Ambient & Cross-Repo QA: Audit local changes as you code (npx jennie qa --ambient) or verify multi-repository microservice breaking changes simultaneously.
• 🎨 Interactive Web Dashboard & 3D Visualizers: Includes Next.js 16 web portal, Framer Motion vertical workflow demo, and interactive 3D system architecture visualizer.

🏗️ System Architecture Workflow

flowchart TD
    subgraph Trigger["1. Execution Trigger"]
        CLI["💻 CLI Review\n(npx @tarunagnihotri534/jennie review)"]
        GHA["🐙 GitHub Actions CI\n(.github/workflows/jennie-review.yml)"]
        WebGUI["🌐 Web GUI Documentation\n(http://localhost:3000)"]
    end

    subgraph Analysis["2. Context & AST Explorer"]
        GitDiff["Git Diff & Modified Files Explorer"]
        ASTParser["AST Call Tree & Dependency Graph"]
        RulesParser["Repo Guidelines Parser (.jennie/rules.md)"]
    end

    subgraph Core["3. Jennie Agent Reasoning Loop"]
        PromptBuilder["Prompt Construction & Context Scaffolder"]
        MCPClient["🔌 Model Context Protocol (MCP Client)"]
        LLM["🧠 LLM Engine (Anthropic / OpenAI / OpenRouter)"]
    end

    subgraph Outputs["4. Output & Reporting"]
        TermReport["💻 ANSI Terminal Report & Security Audit"]
        PRComments["💬 GitHub PR Inline Comments & Summary"]
    end

    Trigger --> Analysis
    Analysis --> Core
    Core --> Outputs

For detailed sequence diagrams and component breakdowns, read the full System Architecture & Workflow Guide.

🚀 Quickstart & Setup

📦 Installation

Install globally via npm or run directly using npx:

# Option 1: Install globally
npm install -g @tarunagnihotri534/jennie

# Option 2: Install as project dependency
npm install @tarunagnihotri534/jennie

1. Local CLI Review & Security Audit

Inspect staged changes or run a codebase-wide security audit:

# Run deep security scan across whole codebase to find sensitive points
npx @tarunagnihotri534/jennie audit

# Review staged git changes in current working directory
npx @tarunagnihotri534/jennie review

# Compare local branch against main with thorough depth
npx @tarunagnihotri534/jennie review --base main --depth thorough --verbose

# Run with custom MCP tool integration enabled
npx @tarunagnihotri534/jennie review --mcp ./mcp-config.json

2. GitHub Actions Integration

Add the workflow file to .github/workflows/jennie-review.yml:

name: Jennie AI Code Review
on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      issues: write

    steps:
      - name: Checkout Code
        uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Setup Node.js
        uses: actions/setup-node@v4
        with:
          node-version: 20

      - name: Run Jennie Review Agent
        env:
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: npx @tarunagnihotri534/jennie review --ci

🔑 Environment Variables

Jennie supports multiple LLM providers. Configure the appropriate keys:

| Environment Variable | Description | Required | Default / Recommended | | :--- | :--- | :--- | :--- | | ANTHROPIC_API_KEY | Anthropic Claude API Key | Yes (if using Anthropic) | Recommended | | OPENAI_API_KEY | OpenAI API Key | Yes (if using OpenAI) | - | | OPENROUTER_API_KEY | OpenRouter Unified API Key | Yes (if using OpenRouter) | - | | CLOUDFLARE_ACCOUNT_ID | Cloudflare Account ID | Yes (if using Cloudflare) | - | | CLOUDFLARE_API_TOKEN | Cloudflare Workers AI Token | Yes (if using Cloudflare) | - | | GITHUB_TOKEN | GitHub Token for PR inline comments | Yes (for GitHub Actions) | secrets.GITHUB_TOKEN | | JENNIE_PROVIDER | Selected LLM Provider | No | anthropic | | JENNIE_MODEL | Override default model name | No | claude-3-7-sonnet-20250219 |

Provider Configuration Examples

Anthropic (Default):

export ANTHROPIC_API_KEY="sk-ant-api03-..."
export JENNIE_MODEL="claude-3-7-sonnet-20250219"

OpenAI:

export OPENAI_API_KEY="sk-proj-..."
export JENNIE_PROVIDER="openai"
export JENNIE_MODEL="gpt-4o"

OpenRouter:

export OPENROUTER_API_KEY="sk-or-v1-..."
export JENNIE_PROVIDER="openrouter"
export JENNIE_MODEL="anthropic/claude-3.5-sonnet"

⚙️ CLI Command Reference

# Run AI code review on git diff
npx @tarunagnihotri534/jennie review [options]

# Run codebase-wide security audit for sensitive points
npx @tarunagnihotri534/jennie audit

| Command / Option | Description | Default | | :--- | :--- | :--- | | audit | Scan whole codebase for secret leaks, unprotected handlers & injection points | - | | review | Run AI code review on git diff or uncommitted changes | - | | --base <branch> | Base target branch for git diff comparison | main | | --ci | Run in CI mode for GitHub Actions output | false | | --depth <level> | Review depth (fast, standard, thorough) | standard | | --max-comments <N> | Maximum number of inline comments posted on PR | 10 | | --mcp <path> | Path to Model Context Protocol config JSON file | - | | --rules <path> | Custom repository rules markdown file | .jennie/rules.md | | --ignore <glob> | File patterns to exclude from review (comma-separated) | dist/**,build/** | | --verbose | Output detailed step-by-step agent reasoning logs | false |

💻 Web Landing Page & Documentation Portal

This repository contains both the Jennie Code Review Agent documentation and its official web marketing portal built with Next.js 16 (App Router) and Tailwind CSS v4.

Local Development

1. Install dependencies:

   npm install

2. Start development server:

   npm run dev

   Open http://localhost:3000 in your browser.

3. Build for production:

   npm run build
   npm run start

4. Run Linter:

   npm run lint

📂 Project Structure

.
├── docs/                        # Complete product documentation
│   ├── setup.md                 # Installation & Quickstart guide
│   ├── action-options.md        # GitHub Action CLI parameters & depth settings
│   ├── ai-provider-config.md    # Multi-provider LLM configuration guide
│   ├── mcp.md                   # Model Context Protocol integration guide
│   ├── rules-files.md           # Custom repository rules (.jennie/rules.md)
│   ├── custom-instructions.md   # Prompt tailoring guide
│   ├── qa-features.md           # Automated QA capabilities
│   ├── ambient-qa.md            # Background daemon setup
│   ├── cross-repo-qa.md         # Multi-repository microservices review
│   └── subagent-tool.md         # Autonomous subagent tools reference
├── src/
│   ├── app/
│   │   ├── docs/                # Interactive documentation browser page (/docs)
│   │   ├── globals.css          # Core design system tokens & Tailwind CSS
│   │   ├── layout.tsx           # Global app shell with ThemeProvider & Navbar/Footer
│   │   └── page.tsx             # Marketing landing page (Hero, Features, QuickStart, FAQ)
│   ├── components/
│   │   ├── layout/              # Navbar, Footer, Mobile Navigation
│   │   ├── marketing/           # Hero, Features, FeatureCard, QuickStart, FAQ
│   │   └── ui/                  # CodeBlock, Badge, ThemeToggle, Stamp, SmoothScroll
│   └── lib/
│       └── content.ts           # Central branding configuration & documentation items
├── public/                      # Static assets & public images
├── package.json                 # Project dependencies & scripts
├── tailwind.config.js / postcss # PostCSS & Tailwind v4 styling setup
└── tsconfig.json                # TypeScript configuration

📚 Complete Documentation Index

For detailed guides, check out the ./docs directory or visit the /docs route on the web application:

• System Architecture & Workflow Guide
• Setup & Quickstart Guide
• GitHub Action Options
• AI Provider Configuration
• Model Context Protocol (MCP) Integration
• Custom Repository Rules
• Prompt Tailoring
• QA Features
• Ambient QA Daemon
• Cross-Repository QA
• Subagent Tools

👤 Author

Tarun Kumar Agnihotri

📄 License

MIT © Tarun Kumar Agnihotri. Distributed under the MIT License. See LICENSE for more information.

Developed with ❤️ for developers who ship fast.