Gemini Workspace Framework

A modular framework for building Gemini-assisted workspaces with consistent structure, reusable skills, and orchestrated workflows.

Quick Start | Documentation | Contributing

Author: Thomas Jamet

---

Why This Exists

AI coding assistants excel at generation but provide no organizational structure. This leads to:

• "Vibe coding" — ad-hoc file creation, no patterns, invisible technical debt
• Context inefficiency — AI agents waste cycles asking "where is this file?"
• Demo vs. reality gap — "build X in 10 minutes" content ignores 6-month maintenance

This framework solves the unsolved problem: long-term maintainability of AI-assisted projects.

Structure scales. Chaos doesn't.

---

Features

• Tiered Architecture: Lite → Standard → Enterprise (matched to project complexity)
• Skills + Workflows: Atomic capabilities + orchestrated sequences
• Built-in Validation: Health monitoring and structure verification
• Upgrade System: Tier upgrades with backup/rollback support

Quick Start

python bootstrap.py -t 2 -n myproject

Project Structure

├── build.py                    # Compiles modules → bootstrap.py
├── config.py                   # Constants, tier definitions
├── core_utils.py               # Exceptions, utilities, validators
├── __main__.py                 # CLI entry point
├── core/                       # Core functionality
│   ├── makefile.py             # Makefile generation
│   └── templates/              # Template generation
├── operations/                 # Workspace operations
│   └── creation.py             # Create, validate, upgrade
└── providers/                  # LLM provider abstraction
    ├── base.py                 # Provider interface
    └── gemini.py               # Gemini implementation

Building

python build.py

Creates bootstrap.py with all modules concatenated into a single distributable file.

---

Questions & Contact

• Discussions: GitHub Discussions
• Issues: GitHub Issues
• Author: LinkedIn

License: MIT