I built a CLI that analyzes pytest failures and proposes fixes automatically. It shows you a diff before applying anything, so you stay in control.
Right now it handles common logic bugs - wrong operators, off-by-one errors, etc. Uses pattern analysis on test output plus an LLM fallback for edge cases.
The goal is to automate the test → fix → validate loop rather than just explaining errors. It re-runs the full suite after each fix to prevent regressions.
Looking for feedback on what kind of failures would be most useful to auto-fix, and beta testers to try it on real projects.
Right now it handles common logic bugs - wrong operators, off-by-one errors, etc. Uses pattern analysis on test output plus an LLM fallback for edge cases.
The goal is to automate the test → fix → validate loop rather than just explaining errors. It re-runs the full suite after each fix to prevent regressions.
Looking for feedback on what kind of failures would be most useful to auto-fix, and beta testers to try it on real projects.