Turning Git Activity into Accurate Developer Work Summaries

I’ve been thinking about a problem with developer work history:

Git records what changed, but not necessarily why it changed.

A commit might show that 12 files were modified, but it doesn’t tell you about the debugging, discarded approaches, decisions, or context that led there.

I ran into this while building a developer work-journal tool around Git activity.

The tempting solution is to let AI reconstruct that missing context, but there’s a problem:

At some point, summarization becomes inference.

If AI sees a refactor, should it explain the likely reason for it? Or should it only report what can actually be proven from the repository?

I’m currently leaning toward keeping factual history separate from AI interpretation and asking the developer when important context is missing.

For people building developer tools or using AI around code:

Where would you draw the line between useful inference and making assumptions about a developer’s work?