ai-sw-pl — Reference-Architecture Run (acc-simulation)

The canonical pipeline, station by station: describe → scaffold (code-blind) → fill (typed) → system tests → deterministic coverage.

reference-architecture pipeline 10 modules · 90 requirements · 644 tests · 100 findings · 34 user-questions

90
requirements (vs. 35 old)
644
system tests
100
findings
34
user-questions
ArtefactWhat it shows
Functional Specification The showcase. Per module, nine-axis requirements with concrete code-derived values; green = deterministic, yellow = interpretive. Firewall at the source.
Findings Behaviour the spec requires that the code does not implement (honest gaps).
User-Questions & Review Calibration / intent values for the customer to supply; ambiguous code to review.
System-Level Test Cases Black-box tests blind to code; every oracle REVIEW REQUIRED.
Coverage Matrix Deterministic requirement→test completeness (the green guarantee).