Docs / source reviewed
PR Scout
Triage a PR backlog, then test-merge and inspect shortlisted diffs
How this project uses Jev
- Input
- GitHub PR titles, descriptions, changed paths and a usage profile, followed by shortlisted diffs and CI status
- Jev decides
- Stage one judges change type, relevance, bug impact and risk; stage two evaluates diff quality, scope and suspicious behavior
- Code executes
- Local scoring rules build a take/consider/skip board; Git checks merge conflicts in a local copy, and optional Ollama fills missing descriptions
Evidence and limitations
MIT source, both decision stages and verdict rules were reviewed without test-merging or calling the model. The board supports human review; a clean merge does not prove behavioral correctness. Some CI failures are excluded from red status by name. Author speed and cost figures were not reproduced, and price comparisons include token-based estimates rather than equal-quality measurements.
This project has not been run independently here. Author-reported results are not independently verified results.
Original sources
- https://github.com/gentslava/pr-scout/blob/cbd0e18532302a69dc87d3807f932ff0eb28885e/README.md ↗
- https://github.com/gentslava/pr-scout/blob/cbd0e18532302a69dc87d3807f932ff0eb28885e/LICENSE ↗
- https://github.com/gentslava/pr-scout/blob/cbd0e18532302a69dc87d3807f932ff0eb28885e/app/main.py ↗
- https://github.com/gentslava/pr-scout/blob/cbd0e18532302a69dc87d3807f932ff0eb28885e/app/scoring.py ↗