Reasoning & Test Intent Engine — the "brain" of autonomous QA.
This crate answers the core questions:
- What changed? (Change analysis)
- What is impacted? (Impact analysis with graph traversal)
- What is NOT impacted? (Equally important — explains WHY)
- What should be tested? (Test intent with confidence tiers)
- What is uncertain? (Ambiguity detection)
- WHY did the system reach these conclusions? (Explanation chains)
Every output includes an explanation_chain — a sequence of
ReasoningStep structs, each with evidence (what data), inference
(what conclusion), and confidence (how sure). This is the
"show your work" layer that makes the system auditable.