🩺 Dockerfile Doctor

Paste your Dockerfile → get a smaller, safer, production-ready version in seconds.

Dockerfile Doctor is an AI-powered optimization tool that automatically rewrites Dockerfiles using best practices. It helps you reduce image size, improve security, fix anti-patterns, and instantly get a clean, optimized Dockerfile.

Built using Kiro, Streamlit, and a modular optimization engine.

---

Try It Locally

streamlit run frontend/app.py

---

Key Features

Intelligent Dockerfile Analysis

Understands your Dockerfile line-by-line and detects:

• Redundant layers

• Inefficient RUN commands

• Missing security hardening

• MLOps-specific stack (PyTorch, TensorFlow, Transformers, vLLM, etc.)

• Build-cache breakers

• Potentially unsafe instructions

⚙️ Automatic Optimization

• Dockerfile Doctor rewrites the file using best-practice patterns:

• Merge and compact RUN layers

• Add non-root user and permissions

• Add recommended packages

• Remove unused instructions

• Auto-detect ML stack and adjust base image

• Improve caching and build reproducibility

🛡️ Zero-Trust Security Score

Each optimized image receives a 0–100 security score based on:

• Root vs non-root user

• Number of unsafe instructions

• Presence of leaked secrets

• Common misconfigurations

• Basic CVE-related checks

Full Diff View

You can see exactly what changed:

--- Original Dockerfile
+++ Optimized Dockerfile
- CMD ["python", "app.py"]
+ RUN useradd -m appuser
+ USER appuser

This makes it easy to review and commit changes safely.

⬇️ One-Click Download

A clean Dockerfile.optimized is generated automatically — ready to push into your repo.

🧩 Home Screen

⚡ Optimization Results

Architecture Overview

dockerfile-doctor/
├── app/
│   ├── analyzer.py        # Dockerfile parser
│   ├── optimizer.py       # Rules engine that rewrites Dockerfile
│   ├── security.py        # Security scoring engine
│   ├── mlops_mode.py      # ML stack detector
│   ├── diff_utils.py      # Diff builder
│   └── models.py          # Pydantic models
├── frontend/
│   └── app.py             # Streamlit UI
└── tests/                 # Unit tests

The backend is modular — each component handles a specific transformation:

• Analyzer → tokenizes and normalizes the Dockerfile

• Optimizer → applies 12+ rewrite rules

• Security Engine → computes the zero-trust score

• MLOps Detector → adjusts optimizations for ML workloads

• Diff Builder → generates unified diffs for UI display

Running Tests

python -m pytest -q

All core components include unit tests to ensure rule reliability.

Local Development

git clone https://github.com/coolhead/dockerfile-doctor.git
cd dockerfile-doctor
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
streamlit run frontend/app.py

Roadmap

• Before/After image size prediction
• BuildKit-powered caching heatmap
• GitHub Action: Auto-PR with optimized Dockerfile
• Trivy-powered CVE scan
• CLI tool (dockerfile-doctor --fix Dockerfile)
• VSCode extension

🤝 Contributing

Contributions are welcome! Feel free to open issues or propose features.

📜 License

MIT License © 2025 Raghavendra S

⭐ Acknowledgements

Built using:

Kiro — AI automation engine

Streamlit — UI framework

Python — core backend

🔗 Links

GitHub Repo: https://github.com/coolhead/dockerfile-doctor

Challenge Submission: https://builder.aws.com/content/361lkwNCmn71f4mdS3HFm4odNzP/building-dockerfile-doctor-an-ai-powered-dockerfile-optimizer-with-kiro-and-streamlit