whetstone_DSL/sprint206_plan.md

Sprint 206 Plan: Anti-False-Green Production Gating

Context

Hard benchmark subset showed production-loop green outcomes where A/B paths reported compile/test failures.

Goals

1. Require explicit compile/test evidence for production green status
2. Reject green status when gate evidence payload is missing/null
3. Emit deterministic false-green diagnostics

Steps

Step 2069: Production evidence completeness gate (10 tests)

Require non-null gate evaluation and evidence packet fields before green.

Step 2070: Compile/test proof binding (10 tests)

Bind green status to successful compile/test proof records.

Step 2071: False-green diagnostic class (8 tests)

Emit deterministic false_green_detected blocked reason.

Step 2072: Production summary hardening (8 tests)

Add explicit evidence_complete and gate_proofs fields.

Step 2073: Sprint 206 Integration Summary (8 tests)

Add editor/src/Sprint206IntegrationSummary.h.

Architecture Gate

• Production loop cannot report overall_ready=true without compile/test evidence packets.