.png&w=256&q=75)
🤓 Latest Submissions
.png&w=828&q=75)
ChangeGuard AI
Problem Software changes rarely affect only the files a developer edits. A change to one module can affect dependent modules, workflows, inventory logic, payments, APIs, and other behavior. Developers therefore need to understand the broader impact of a change and determine whether existing tests actually cover the affected behavior. Manual impact analysis and regression-test selection can be time-consuming and can leave hidden defects undetected. Solution ChangeGuard AI is an AI-assisted developer workflow that analyzes a code change, maps potentially affected parts of the repository, identifies regression-test gaps, generates targeted regression tests, executes them, discovers defects, and validates fixes. The prototype uses a controlled e-commerce backend repository, change-impact-demo, containing users, authentication, products, carts, payments, orders, notifications, and automated tests. For the demonstration, a discount-code feature was introduced into the payment and order flow. The workflow identified affected areas and six initial regression gaps. Generated tests exposed a stock-restoration defect after failed payment. After the fix, deeper impact analysis identified another gap in cancelOrder(), where cancelled orders did not restore inventory. Additional regression tests exposed the issue, the fix was applied, and the final suite passed 89/89 tests across 10 suites. Target Users The primary users are software developers working on existing codebases. Code reviewers and engineering teams can also use the resulting impact and validation information when reviewing changes. Developers remain responsible for deciding whether a change should be merged. Interaction A developer provides a change description and repository context. ChangeGuard analyzes the change, presents impacted files and test gaps, generates and runs targeted regression tests, reports discovered defects, and shows the final validation result. Creative and Unique Aspect Instead of focusing on generating new application code, ChangeGuard AI focuses on the consequences of code changes. Its workflow connects impact analysis, test-gap discovery, regression generation, defect discovery, repair, and final validation into one agentic development loop. Effectiveness The prototype demonstrated that this workflow can uncover defects that were not detected by the existing test suite and can increase regression coverage while providing evidence that the repaired code passes the complete test suite.
27 Sep 2026
