Show HN: Causari – Content-addressable ledger for AI agent code edits (causari.dev)

Causari has introduced a groundbreaking content-addressable ledger designed for tracking AI agent code edits, providing a solution to the common problem of obscured code provenance. Unlike traditional tools that focus solely on code or agent behavior, Causari captures every interaction—including prompts, model responses, and reasoning—as immutable events. This allows developers to trace back how specific lines of code were conceived and altered, effectively linking code changes to their originating intent, and preventing the loss of context when reverting agent decisions. Significantly, Causari's capture engine operates without requiring the agent's cooperation, thereby enhancing transparency and accountability in AI-driven development. It combines a local proxy for LLM interactions with a passive filesystem recorder, creating a causal graph that reveals both upstream influences and downstream impacts of code changes. The result is a system that not only tracks code edits but also generates signed skills based on the agent's past experiences, fostering trust in the AI's capabilities. This innovation is poised to dramatically enhance code retrieval, error resolution, and overall project management, allowing engineers to maintain high standards of reliability and traceability in AI-influenced codebases.