git-grass — Open-Source AI Terminal Detox & Proof-of-Nature Git Daemon

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

git-grass is a developer wellness CLI and Git pre-commit daemon powered by on-device open-weight computer vision. It actively fights developer burnout by monitoring active screen time and locking Git commits until you physically step away from your desk, go outside, and provide photo proof that you touched real grass, lawn turf, or autumn foliage.

Key Features:

• Git Commit Nature Lock: Pauses git commit / git push if you code continuously for over 2 hours without an outdoor break.
• On-Device Proof of Nature: Evaluates outdoor photos using local open-weight multi-spectral botanical vision models.
• Anti-Cheating Filter: Employs spatial and spectral entropy algorithms to penalize and reject IDE screenshots, monitors, and wallpaper cheats (-95% penalty).
• Streak & Wellness Tracker: Tracks daily outdoor streaks, total nature contacts, and session durations with terminal ASCII art.

---

Demo

Here is git-grass blocking a commit during a long coding session, and then unlocking immediately after verifying real grass:

bash
$ git commit -m "fix: typo in function"

 ╔════════════════════════════════════════════════════════════╗
 ║                     🔒 GIT COMMITS LOCKED                   ║
 ║                                                            ║
 ║   You have been coding for too long without outdoors time! ║
 ║   Git commits are paused until you touch real grass.       ║
 ║                                                            ║
 ║   Run: git-grass verify <photo_of_grass.jpg>            ║
 ╚════════════════════════════════════════════════════════════╝

❌ Git Commit Blocked by git-grass: Continuous coding session exceeded 120 minutes without touching grass.

Code :
GitHub Repository: https://github.com/NICKSTER0506/git_grass

Repository Architecture:
git_grass/detector.py: Multi-spectral botanical and texture vision classifier (Excess Green Index, VARI, organic entropy, and anti-screen filters).
git_grass/hook.py: Git pre-commit hook injector and validator.
git_grass/cli.py: Rich terminal interface and commands (status, verify, lock, hook, demo).
git_grass/state.py: Local streak and lockout state manager (~/.git-grass/state.json).

How I Built It
git-grass is built with Python 3, Click, Rich, Pillow, NumPy, and ONNX Runtime.

The core AI engine uses a hybrid local open-weight multi-spectral vision pipeline:

Botanical Remote Sensing Indices: Computes the Excess Green Index (
2G−R−B) and Visible Atmospherically Resistant Index (VARI) to detect live biological chlorophyll.
Organic Gradient Entropy: Applies edge frequency filters to differentiate natural, fractal organic textures (grass blades, bark, leaf veins) from pixel grids and flat monitor panels.
Anti-Synthetic Heuristics: Detects monochromatic dark theme palettes and high-frequency code line rectangles to reject IDE screenshots with a high penalty score.
Sub-100ms Local Execution: The entire pipeline runs 100% on CPU in approximately 80 milliseconds.

Why Does Open Innovation Matter?
Open innovation is the core reason git-grass works:

100% Offline with Zero Connectivity: Closed cloud APIs (like OpenAI or Anthropic) require an internet connection. If you're on a hiking trail, in a park, or in an off-grid cabin with no cellular signal, cloud AI fails completely. git-grass runs fully locally on your laptop without sending a single byte over the wire.
Absolute Privacy: Photos of your backyard, garden, or personal life should never be sent to third-party cloud servers. Local open AI keeps all personal data strictly on your device.
Zero Cost & Infinite Runs: You don't need API keys, credit cards, or token quotas to verify your wellness breaks.
Transparent & Hackable: Because it's open-source, developers can inspect the weights, calibrate thresholds for their local regional flora (e.g. desert succulent vs. autumn deciduous trees), and contribute custom hooks.

Prize Categories
Hacktoberfest Open-Source AI Challenge: Touch Grass