What problem does it solve? Hard bugs and flaky performance regressions resist casual code reading; without a reproducible pass/fail signal, debugging devolves into guessing. This Skill enforces a disciplined six-phase diagnosis loop that builds a tight reproduction command before any hypothesis is tested. ## Core Features & Use Cases - Feedback Loop Construction: Builds a red-capable, deterministic, fast reproduction command using failing tests, curl scripts, CLI fixtures, Playwright browser scripts, trace replay, fuzz loops, or git bisect harnesses. - Ranked Hypothesis Testing: Generates 3-5 falsifiable hypotheses, instruments with tagged debug logs, and changes one variable at a time to isolate the root cause. - Regression Lockdown: Converts the minimized repro into a failing regression test at a correct seam, verifies the fix, and removes all instrumentation during cleanup. - Use Case: A user reports that an export endpoint intermittently throws a 500 error. The Skill drives creation of a curl-based loop that reproduces the failure, minimizes the input, tests ranked hypotheses, and lands a fix with a regression test. ## Quick Start Ask the agent to diagnose the failing export endpoint using the diagnosing-bugs skill and build a reproduction loop first.
