# Sprint 149 Plan: Generic AST-to-Text Regeneration and Diff Stability

## Context

Sprint 149 delivers shared regeneration orchestration with language-specific
style policies provided by adapters, minimizing non-semantic churn.

---

## Goals

1. Implement language-neutral AST-to-text regeneration orchestration
2. Plug in adapter style/include/order policies without core branching
3. Minimize semantic-noise diffs across supported languages
4. Publish deterministic regeneration decisions and quality evidence

---

## Steps

### Step 1729: Generic regeneration policy schema (10 tests)
### Step 1730: Adapter style policy binding model (10 tests)
### Step 1731: Include/import ordering normalization model (10 tests)
### Step 1732: Semantic-noise suppression comparator model (10 tests)
### Step 1733: `whetstone_regenerate_text_generic` MCP tool (8 tests)
### Step 1734: `whetstone_preview_regeneration_diff_generic` MCP tool (8 tests)
### Step 1735: `whetstone_get_regeneration_trace_generic` MCP tool (8 tests)
### Step 1736: Cross-language regeneration quality packet model (8 tests)
### Step 1737: Regeneration stability report artifact (8 tests)
### Step 1738: Sprint 149 integration summary + regression (8 tests)

---

## Architecture Gate

- Regeneration orchestration must remain language-neutral
- Adapter policies must be explicit and versioned
- Diff stability metrics must be reproducible and comparable across languages
- Max 600 lines per header