What problem does it solve?
This Skill solves the problem of wasted engineering effort and incorrect bug fixes caused by unverified root cause diagnoses and unvalidated prescribed fixes from task briefs, which often lead to changes that mask real issues or introduce new failures.
Core Features & Use Cases
- Dual Hypothesis Verification: Validates both the stated root cause and the prescribed fix/transform before any code is written, preventing changes based on incorrect assumptions about data semantics or lever direction.
- Scale Mismatch Detection: Identifies hidden scale mismatches (aggregate vs summed, per-unit vs total) that cause constant or all-zero metrics by instrumenting both sides of binding comparisons.
- Shared Component Falsification: Rules out global shared component breakage by sampling 3-5 sibling consumers before building unnecessary workarounds.
- Proxy Input Discipline: Flags findings based on proxy or stand-in data as provisional, tracks required real input artifacts, and triggers recomputation when real data becomes available.
- Newly Reachable Path Testing: Runs end-to-end tests for code paths that become reachable only after a guard is removed, catching downstream failures that unit tests miss.
- Use Case: When a task brief prescribes differencing a cumulative distribution to fix an all-zero metric, use this Skill to first verify the distribution is actually cumulative (not bell-shaped per-year) before applying the transform, avoiding negative distributions that corrupt economic analysis.
Quick Start
Use the systematic-debugging-extras skill alongside systematic-debugging to verify the root cause and prescribed fix for the all-zero Q3 marketing metric before implementing any changes.
