Diagnose hard bugs through a six-phase loop from reproduction to regression-testing.
npx skills add https://github.com/hgxszhj/mattpocock_skills --skill diagnose-hgxszhj
Please help me install this Agent Skill. Skill: diagnose Source: https://github.com/hgxszhj/mattpocock_skills/tree/main/skills/engineering/diagnose Command: npx skills add https://github.com/hgxszhj/mattpocock_skills --skill diagnose-hgxszhj
Disciplined diagnosis loop for hard bugs and performance regressions by guiding you from reproduction → minimise → hypothesise → instrument → fix → regression-test.
Run the six-phase diagnosis loop on the active bug report to reproduce, minimise, hypothesise, instrument, fix, and regression-test.
High-intent search queries and answers about installing and using this skill.
Systematic debugging for hard bugs requires a deterministic feedback loop, instrumentation for signals, and explicit progression criteria to validate hypotheses. It applies to crashes, performance regressions, and ambiguous failures across projects.
Debug a performance regression by applying the six-phase diagnosis loop: reproduce the drop, minimise the trigger, hypothesise the cause, instrument for deterministic signals, apply the fix, and regression-test the result.
Minimise a bug reproduction by isolating the deterministic trigger and stripping extraneous variables. The debugging loop uses this minimised case to build fast feedback loops, select deterministic signals, and document testable hypotheses.
Instrumentation is needed during software debugging to capture deterministic signals that validate or invalidate documented hypotheses. It provides the explicit feedback required before progressing to the fix and regression-testing phases.
Yes, this debugging workflow applies to software debugging tasks across projects with a defined glossary and ADRs. These artifacts provide the contextual boundaries needed for accurate reproduction and regression-testing.
The disciplined diagnosis loop for ambiguous failures requires a deterministic feedback loop to function correctly. Without deterministic signals and explicit progression criteria, validating hypotheses and fixes for unpredictable failures becomes unreliable.
jacob-balslevExtract durable project knowledge from code, docs, issues, and reports.
rabbicseAnalyze errors, stack traces, and logs to isolate software failure root causes.
KnoWhizMaintain persistent Playwright handles across debugging sessions for web and Electron apps.
wachawoDiagnose software bugs and performance regressions through a six-phase methodology.
