# Sprint 152 Plan: Language-Aware Constructive Loop Orchestrator

## Context

Sprint 152 composes neutral core operations and toolchain providers into a
single constructive loop that supports multiple language execution plans.

---

## Goals

1. Orchestrate edit -> sync -> merge -> build/test -> feedback generically
2. Select language-specific execution plans from declared capabilities
3. Bound loop latency with deterministic retries and failure classes
4. Expose one-call loop APIs without language-hardcoded branches

---

## Steps

### Step 1759: Constructive loop orchestration schema (10 tests)
### Step 1760: Language execution plan resolver model (10 tests)
### Step 1761: Loop retry and timeout policy model (10 tests)
### Step 1762: Feedback aggregation and prioritization model (10 tests)
### Step 1763: `whetstone_run_constructive_step` MCP tool (8 tests)
### Step 1764: `whetstone_get_constructive_status` MCP tool (8 tests)
### Step 1765: `whetstone_run_constructive_loop` MCP tool (8 tests)
### Step 1766: Loop readiness score packet model (8 tests)
### Step 1767: Multi-language constructive loop report artifact (8 tests)
### Step 1768: Sprint 152 integration summary + regression (8 tests)

---

## Architecture Gate

- Orchestration must consume capability matrix, not language conditionals
- Retry/timeout behavior must be deterministic and bounded
- Loop status must expose structured stage-level evidence
- Max 600 lines per header