Converts bug reports into verified fixes with reproducible tests.
npx skills add https://github.com/ai-pivot/xbot --skill debug-ai-pivot
Please help me install this Agent Skill. Skill: debug Source: https://github.com/ai-pivot/xbot/tree/main/tools/embed_skills/debug Command: npx skills add https://github.com/ai-pivot/xbot --skill debug-ai-pivot
Debug turns unclear bug reports into a reproducible, verifiable fix plan so you can stop guessing and confirm correctness.
Ask the skill to debug a reported issue by requesting: the reproduction steps, expected vs actual behavior, any failing command or test output, and a success criterion for the fix.
High-intent search queries and answers about installing and using this skill.
To reproduce a bug from a vague crash report, you need to clarify the observed versus expected behavior and define concrete success criteria. This approach converts unclear issues into a verifiable fix plan by creating the smallest failing automated check possible to confirm the crash.
The best way to fix a regression is to scope changes strictly to the verified cause after creating a failing test. By re-running the reproduction and nearby validations post-fix, you confirm the regression is resolved without introducing new issues.
To debug a flaky test in your CI workflow, turn the reported issue into a reproducible failure by defining expected behavior and adding a failing automated check. Once the flaky behavior is consistently reproduced, apply fixes and re-run validations to verify correctness.
Yes, test-first debugging works for incorrect outputs and regressions by requiring a failing automated check to exist before any code changes are applied. This ensures the incorrect output is reliably reproduced and verified against defined success criteria.
Your bug fix requires a success criterion to stop guessing and confirm correctness. Defining expected behavior upfront ensures the fix is scoped to the verified cause and validated by re-running the reproduction test.
To investigate a failing behavior, you need to provide reproduction steps, the expected versus actual behavior, any failing command or test output, and a success criterion for the fix. This evidence drives the test-first debugging process.
pcasarettoReproduce reported software issues and classify them as bugs, environmental problems, or user errors.
jpcaparasConvert documented heuristics into deterministic validation scripts and CI checks.
mikkeymooAutomate targeted code refactoring with automated tests to preserve behavior.
ananthamIntegrate static analysis, reproduction, classification, and archaeology tools for bug investigation workflows.
