The CHAC Workbench

Protocol Version: CHAC v14.0 (Internal)

Welcome to the CHAC Workbench, the first and primary implementation of the Compensatory Human-AI Collaboration (CHAC) framework. This repository contains the complete, living history of the framework's evolution and serves as a template for bootstrapping new CHAC-based research projects.

---

🚀 How to Use This Framework

1. Creating Your Own Experiment (Required First Step)

Why is this important? The core philosophy of the CHAC framework is compensation. The AI is designed to adapt to and compensate for a specific user's cognitive and emotional profile. To unlock the true power of this framework, you must create your own experiment and provide the AI with a description of the user it is meant to collaborate with.

How to do it:

1. Duplicate the template: Copy the entire experiments/_template/ directory.
2. Rename it: Give it a meaningful name for your project, e.g., experiments/my_research/.
3. Customize your profile: This is the most critical step. Follow the chac_toolkit/CHAC_Profile_Co-Creation_Protocol.md to collaboratively generate a high-fidelity profile.md file inside your new directory.
4. Activate your experiment: Open the config.json file in the root directory and change the active_experiment to point to your new directory name.

2. Starting a Work Session (The v11.0 "Genesis" Boot)

The v11.0 framework introduces a robust, single-prompt boot process that is both simpler and more reliable.

1. Launch your AI CLI in the project's root directory.
2. Execute the Genesis Prompt:
   • Copy the entire content of the file chac_toolkit/00_CHAC_Genesis.md.
   • Paste it as your first and only initialization prompt to the AI and send.

The AI will now automatically read and internalize the entire boot sequence (00_ to 05_) in the correct order. Upon completion, it will output a final confirmation line: CHAC framework v10.1 initialized. Standing by. and will then await your first task-oriented prompt.

---

🗺️ Repository Guide

Core Directories

• experiments/: Contains one or more "experiment packages." Each package is a self-contained configuration for the CHAC framework.
• chac_toolkit/: The heart of the framework. Contains all the core protocols that define the AI's identity and behavior. See the chac_toolkit/README.md for a detailed breakdown of the boot sequence and all protocol files.
• case-study/: Contains all empirical case study data and analysis for this project.
• build/: Contains the output of the CHAC_SOP_Modular_Prompt_Builder.md process, such as standalone prompts.

Configuration

• config.json: A simple configuration file that tells the framework which experiment package to load.
  • active_experiment: The name of the directory in experiments/ to use.