What problem does it solve?
This Skill prevents guess-and-check debugging by forcing you to form explicit, evidence-based hypotheses before attempting any fix, reducing time spent patching symptoms and increasing root-cause clarity.
Core Features & Use Cases
- Hypothesis-first loop: Requires 3 ordered hypotheses and a targeted test for the most likely one.
- Deterministic feedback loops: Focuses on building a fast pass/fail signal (tests, scripts, traces, or stress loops) before changing code.
- Evidence recording and iteration: Tracks which hypotheses are confirmed or ruled out, then iterates until the root cause is found.
Use cases include investigating failing tests, runtime errors, broken behavior, and performance regressions—especially when outcomes are unclear or intermittent.
Quick Start
Ask an agent to debug a failing test using the systematic-debugging loop by first constructing a deterministic feedback loop, reproducing the failure, writing three hypotheses, testing H1, recording evidence, and only then implementing a confidence-backed fix with a regression test.