What vani demonstrated

Built and verified a full legacy-service repair for a car-maintenance codebase: fixed the wear calculation, handled missing service readings without crashing, corrected the fleet report’s averaging, and preserved the 15000 km / 80% business rules. Also added the missing regression test, documented the agent review in NOTES.md, and produced a data-driven breakdown-risk analysis that ranks cars by the factors the fleet history actually supports.

• Leakage-Free Bug Diagnosis
• Verified AI Agent Auditing
• Data-Driven Risk Analysis
• Clear Technical Communication