What problem does it solve? Hard bugs and flaky performance regressions resist casual code reading; without a tight, reproducible pass/fail signal, debugging devolves into guessing. This Skill enforces a disciplined diagnosis loop so the root cause is found and locked down with a regression test. ## Core Features & Use Cases - Feedback-loop construction: Builds a tight, deterministic, red-capable reproduction command using failing tests, curl scripts, headless browsers, replayed traces, fuzz loops, or git bisect harnesses. - Structured six-phase process: Reproduce and minimise, generate 3-5 ranked falsifiable hypotheses, instrument with tagged debug logs, fix with a regression test, then clean up and post-mortem. - Human-in-the-loop fallback: Ships a bash template (scripts/hitl-loop.template.sh) that drives a human through manual reproduction steps and captures observations for the agent. - Use Case: A user reports that the export endpoint intermittently returns empty files. The Skill builds a curl-based loop that reproduces the failure, minimises the scenario, tests ranked hypotheses one variable at a time, and lands a fix guarded by a regression test. ## Quick Start Ask the agent to diagnose the bug where the export button intermittently throws an error, and have it build a reproduction loop before proposing any fix.
