whogitit

whogitit is a powerful tool for tracking AI-generated code with precision and clarity, enabling developers to identify which lines of code were written by AI, which were modified by humans, and the prompts that generated them. This project is designed to enhance collaboration and accountability within development teams by providing comprehensive insights into code contributions.

Key Features

• Line-Level Attribution: Gain visibility into whether each line of code is AI-generated, human-modified, or original, allowing for detailed understanding of contributions.
• Prompt Preservation: Retain the prompts that led to the creation of each code segment, facilitating better context and traceability.
• Three-Way Diff Analysis: Ensure accurate attribution by capturing snapshots of file states before, during, and after modifications.
• Git-Native Storage: Utilize git notes for storing attribution data that integrates seamlessly with Git workflows.
• Claude Code Integration: Automatically capture AI additions through integration hooks, simplifying the process for developers.
• GitHub Action Support: Automatically generate pull request comments that summarize AI attributions, keeping all team members informed.
• Privacy Protection: Automatically redact sensitive information such as API keys and passwords, enhancing security.
• Configurable Data Retention Policies: Set retention limits for compliance needs and automatically purge outdated data.
• Audit Logging: Track and log changes, exports, and configuration adjustments for transparency.
• Export Capabilities: Easily bulk export attribution data in formats like JSON or CSV for analysis and reporting.

Usage Examples

Show AI Attribution for Lines

whogitit blame src/main.rs

View Attribution for a Specific Commit

whogitit show HEAD

Generate Summary for a Commit Range

whogitit summary --base main --head HEAD

Manage Data Retention Settings

whogitit retention config

Additional Insights

Data Flow

The process starts with the writing and editing of code using Claude Code tools, capturing changes and prompting continuously. The captured data is processed to maintain an accurate record of contributions, leading to high-quality attribution stored within Git notes.

Privacy Measures

Prompts are scrutinized for sensitive information, which is redacted prior to storage to ensure that private data remains secure and compliant with best practices.

whogitit thus stands as an essential component in modern software development, combining the benefits of AI-assisted coding with comprehensive tracking and accountability features.