What problem does it solve? Deciding what to test for a new feature is high-leverage but error-prone: requirements-only passes miss implementation risks, and case-level lists bury the coverage shape. This Skill turns a feature into a code-grounded test scope with risk-ordered test groups, target suites, scope boundaries, and an acceptance gate. ## Core Features & Use Cases - Code-grounded scoping: Inspects the implementation, Bronze-to-Gold data flow, consumer contracts, and existing tests to expose gaps, risks, and spec disagreements before writing any test plan. - Shape-aware strategy: Adapts the gate to the feature type — differential parity for ports and consolidations, security fail-closed matrices for platform replacements, reusable per-wave gates for rolling migrations, and correctness ordering for new capabilities. - Lean actionable output: Produces a compact, risk-ordered plan (roughly six bullets) with suite assignments, an out-of-scope boundary, and an acceptance gate, ready to file as a linked GitHub test subtask. - Use Case: Given a GitHub issue for a metrics consolidation feature, the Skill reads the branch code, discovers the new counters must reproduce retired sibling keys, and files a scope whose headline gate is a differential with a sum-equals-parts invariant. ## Quick Start Ask the assistant to scope the tests for Insight issue #1538 using the scope-feature-tests skill and draft the resulting test subtask for review.
