WIP: stage all uncommitted work — sprints 46-221, graduation headers, specialist fleet, test steps 909-1988

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

sprint56_plan.md

@@ -0,0 +1,36 @@
+# Sprint 56 Plan: Data and Query Family (SQL Dialects + Dataflow)
+
+## Context
+
+SQL/TSQL/PostgreSQL/MySQL semantics and transaction behavior are central in real
+legacy systems. Sprint 56 adds explicit query and transaction intent modeling.
+
+---
+
+## Goals
+
+1. Normalize SQL dialect semantics into IR query model
+2. Preserve transaction and isolation intent across projections
+3. Support query->application-layer and application->query projections
+4. Add risk gates for semantic shifts in joins/null/aggregation behavior
+
+---
+
+## Steps
+
+### Step 789: SQL canonical query IR layer (12 tests)
+### Step 790: PostgreSQL lowering/raising adapters (10 tests)
+### Step 791: T-SQL lowering/raising adapters (10 tests)
+### Step 792: MySQL lowering/raising adapters (10 tests)
+### Step 793: Transaction/isolation semantics packet model (10 tests)
+### Step 794: Query behavior equivalence runner (set-based diff checks) (10 tests)
+### Step 795: Null/join/aggregation divergence classifier (8 tests)
+### Step 796: `whetstone_transpile_query_family` MCP tool (8 tests)
+### Step 797: Data family acceptance report (8 tests)
+### Step 798: Sprint 56 integration summary + regression (8 tests)
+
+---
+
+## Gate Requirement
+
+- Query projections must include semantic diff evidence for representative datasets.