What problem does it solve? Fixes to the AI Measurement pipeline often ship without tests that would have caught the original regression, allowing perimeter, topology, state, and report-rendering bugs to silently recur. This Skill mandates that every pipeline change includes regression tests validated against canonical baselines like the Fonsica roof benchmark. ## Core Features & Use Cases - Baseline Enforcement: Asserts Fonsica canonical values (3077 sqft area, ~264 LF perimeter, ~14 facets, 6:12 pitch, non-zero ridge and valley lengths) on every relevant change. - Gate-by-Gate Assertions: Validates perimeter shape checks, topology phase execution, route provenance, result_state normalization, and customer_report_ready transitions individually. - Report Rendering Guards: Ensures rejected-geometry rows never render as customer-ready reports and that rejected_only diagrams still show the perimeter overlay. - Use Case: When fixing a bug where topology ran despite a failed perimeter shape check, use this Skill to generate a Deno test asserting phase3C/phase3D never execute when shape_passed is false. ## Quick Start Ask the AI to add regression tests for a measurement pipeline fix, for example: create a Fonsica perimeter shape regression test for the start-ai-measurement function.
