Causari CLI tracks AI coding agent intent and causality alongside Git

A developer has released Causari, an open-source local command-line tool designed to record why AI coding agents make code changes, not just what they changed. As AI agents grow capable of modifying large codebases rapidly, standard Git logs fail to capture which prompts, models, or reasoning led to specific lines of code. Causari logs agent actions as causal events, storing prompts, file reads and writes, diffs, model details, token costs, and relationships between actions. It operates either as a local LLM proxy that intercepts agent requests or as a filesystem watcher that tracks file changes independently. The tool is positioned as a complement to Git rather than a replacement, aiming to provide intent provenance for AI-driven development workflows.