A premium, high-fidelity security, vulnerability, and technical debt analysis dashboard for modern engineering and security teams.
DebtRadar visualizes security risks, code-complexity propagation pathways, and software engineering technical debt. It maps scanned issues directly to standard industry security categories (OWASP Top 10, CWE definitions) and leverages HuggingFace Mistral LLM inference endpoints to generate non-technical summaries for executives and dynamic, click-to-deploy code patches (Autofixes).
- Dynamic Simulation Canvas: High-fidelity, force-directed network showing modules as nodes. Node size represents code blast radius and complexity.
- Risk Heat Zones: Color-gradient indicators dynamically display severity zones (Critical, High, Medium, Low).
- Depth & Elevation: Custom SVG drop-shadow filter layers render nodes with high contrast against the luxurious light sand-beige canvas.
- Unified Health Indicators: High-contrast, bold-bordered dashboard elements highlighting repository health scores:
- Repo Security Score: Comprehensive security health standing.
- Trust Score & Deployment Confidence: Predicts production deployment stability and compliance index.
- Collapse Risk & Exploitability: Highlights cascade failure risks and public exploit surfaces.
- Business Translation Card: Breaks down obscure codebase flaws into non-technical language (Customer Impact, Operational Risk, Action Required).
- Automatic Code Autofixer: Generate proposed remediation patches for vulnerable sections on-demand using mistral models, featuring live visual diff-patches (
git diffrepresentation) and click-to-approve code updates. - Intelligent Deduplication: Automatically deduplicates similar backend-generated forecast insights for an clean, uncluttered user interface.
- Zero-White Design Philosophy: Completely styled with a warm-neutral glassmorphism theme (
bg-[#f5efe7]backdrops, solid sandbg-[#efe8de]card panels, cocoa-brown highlights). - Micro-Interactive Physics: Includes elastic buttons (
active:scale-95), float offsets (hover:-translate-y-0.5), and rotating map indicators on hover.
graph TD
A[Github Repository] -->|Scanned Symbol AST| B(Lume Analysis Pipeline)
B -->|OWASP & CWE Mapping| C[Supabase Postgres DB]
C -->|Next.js REST API| D[Frontend React Clients]
D -->|Interaction Selection| E[D3 Force Graph Canvas]
D -->|Trigger Autofix| F[HF mistral-7b API]
F -->|Unified Diff Patch| D
- Framework: Next.js 14 (App Router, React Suspense, Dynamic Routing)
- Data Visualization: D3.js (Force Simulations, Interactive Canvas Zoom, Drag Vectors)
- Styling: Tailwind CSS (Premium Warm-Neutral Glassmorphic Color Palette)
- Database & Auth: Supabase Server Client (Supabase Service Client serverless integrations)
- LLM Engine: HuggingFace Inference API (Mistral-7B-Instruct / Zephyr-7B LLM orchestration)
- Node.js (v18.x or above)
- npm / pnpm / yarn
- Supabase Account & HuggingFace API Key
-
Clone the repository:
git clone https://github.com/your-username/Lume.git cd Lume -
Install all dependencies:
npm install # or pnpm install -
Configure Environment Variables: Create a
.env.localfile in the root directory and add:NEXT_PUBLIC_SUPABASE_URL=your-supabase-project-url NEXT_PUBLIC_SUPABASE_ANON_KEY=your-supabase-anon-key SUPABASE_SERVICE_ROLE_KEY=your-supabase-service-role-key HUGGINGFACE_API_KEY=your-hf-api-key GITHUB_PAT=your-github-personal-access-token
-
Launch Local Dev Server:
npm run dev # or pnpm devOpen http://localhost:3000 to view the premium dashboard.
-
Build for Production:
npm run build
Distributed under the MIT License. See LICENSE for more information.