What problem does it solve? Hard bugs and performance regressions resist ad-hoc debugging: without a reproducible signal, fixes are guesses. This Skill enforces a disciplined diagnosis loop that builds a tight pass/fail feedback signal first, then drives reproduction, hypothesis testing, instrumentation, fixing, and cleanup. ## Core Features & Use Cases - Feedback loop construction: Builds a tight, deterministic, red-capable loop (failing test, curl script, CLI invocation, headless browser, replayed trace, bisection harness, or human-in-the-loop script) before any fixing begins. - Structured hypothesis testing: Generates 3-5 ranked, falsifiable hypotheses, then instruments with tagged debug logs or profilers, changing one variable at a time. - Regression and cleanup discipline: Writes a regression test at a correct seam before the fix, removes all tagged instrumentation, and records the confirmed root cause in the commit message. - Use Case: A user reports an intermittent export failure in production. The Skill replays a captured request through the code path, minimizes the repro, tests ranked hypotheses, lands a fix with a regression test, and hands architectural findings off for follow-up. ## Quick Start Ask the agent to diagnose the bug where the export button throws an intermittent error and have it build a reproduction loop first.
