If AI Is Doing the Investigation, Version the Investigation (wingedpig.com)

A developer has introduced an innovative open-source development environment called Trellis, created to address the persistent issue of preserving contextual knowledge during software investigations. Many developers, including the author, rely on AI tools like Claude Code to troubleshoot bugs, but findings and thought processes from these sessions often disappear once the investigation is over. Trellis solves this problem by allowing users to save comprehensive "Cases" alongside their code commits, which include a summary, AI session transcripts, and other relevant data, ensuring that critical insights are not lost over time. This development is significant for the AI/ML community as it emphasizes the need for AI-assisted workflows to provide long-term value by maintaining a historical context for software changes. By embedding knowledge directly within the code repository, Trellis facilitates better collaboration and code review, allowing developers to understand not just what was changed, but why it was changed—making it easier to evolve projects in the future. Additionally, Trellis ensures that AI can safely interact with production environments, focusing on investigation without risking system integrity, thereby setting a precedent for how artificial intelligence should be integrated into software development practices.