forensic-deepdive
An open-source, multi-language static-analysis engine that turns a codebase and its Git history into a queryable knowledge graph with explicit confidence levels for every relationship.
forensic-deepdive is an Apache-2.0 code-intelligence and software-forensics engine built for both engineers and AI coding agents. It analyses a repository as a living system: source structure, dependency relationships, architectural seams, and the Git history that explains how those pieces evolved.
Confidence-aware code intelligence
The graph does not pretend every relationship is equally certain. Each edge carries an explicit confidence taxonomy - EXTRACTED, INFERRED, or AMBIGUOUS - so downstream tools can distinguish facts directly observed in code from reasoned hypotheses that deserve review.
Git history as architecture data
Repository history is modelled alongside source structure rather than treated as a separate report. That makes it possible to surface hotspots, co-change patterns, ownership signals, and the sequence of decisions behind present-day architecture.
Built for agents and humans
The engine produces six output surfaces, including queryable graph data and Obsidian-ready documentation. The goal is to give coding agents a grounded map of an unfamiliar codebase while giving engineers documentation they can inspect and challenge.
Disciplined delivery
Version 0.8 was shipped through an ADR-driven roadmap with more than 90 logged decisions, keeping the architecture auditable as the analysis surface expanded across languages and outputs.