Stephen Mc Neill Killeen@ITz-_-Stephen

Software teams often rely on coverage percentages and passing test suites as signals of quality, but those signals can hide important untested behaviour. Error paths, boundary conditions, validation logic, alternate branches, and newly added functions may remain uncovered even when every existing test passes. Finding those gaps manually is slow, especially in unfamiliar codebases, and writing focused tests for each missing path takes additional time. Test Gap Finder is a developer tool that turns that manual process into an analysis pipeline. A developer provides either a trusted local repository path or a public GitHub repository URL. The application scans Python source and test files, extracts functions, branches, raises, calls and existing test behaviour, then compares the code against the current tests to identify likely gaps. Each finding includes a confidence level and an explanation so developers can distinguish stronger findings from conservative heuristics. For trusted local repositories, Test Gap Finder converts structured gap plans into executable pytest tests, validates generated syntax and collection, runs the original and generated suites, and reports before/after coverage. In our controlled demo repository, the tool finds gaps across three functions, generates five targeted tests, increases the passing suite from four original tests to nine total passing tests, and raises coverage from 65% to 92% — a 27 percentage-point improvement. For external GitHub repositories, the product deliberately switches to static-analysis-only mode. Repositories are shallow-cloned into a temporary directory, analyzed without importing or executing their code, and then removed. This gives developers useful findings on real projects without treating downloaded code as trusted. The Angular frontend presents repository history, detected gaps, generated tests, test results and coverage impact in one dashboard. The backend uses FastAPI and an AST-based Python analyzer, with an AI planning provider and deterministic fallback. The result is a practical testing assistant that focuses on *what is missing*, not simply whether the current suite passes.

💻 Online Hackathon | ⏳ Build: 48 hours | 📅 September 25–27, 2026 The official IBM Bob 2.0 hackathon. Build what's next in AI-assisted development, alongside the developers and experts shaping it. ⭐ Get help from experts throughout the event 🤝 Join a team or build solo 💰 Prize pool: $12,000 🧑🏻‍💻 Register before the Kick-Off Stream to start building with Bob