[Change] specfact-code-review module scaffold with ReviewFinding/ReviewReport models

[djm81]

Change: specfact-code-review Module Scaffold with Pydantic Models

Why

The current coding automation pipeline runs a generic codex review pass but has no structured scoring, no persistent quality ledger, and no contract-bound enforcement gates. A dedicated nold-ai/specfact-code-review installable module closes these gaps by providing a governed entry point for all code-review subcommands under specfact code review.

This change establishes the foundation: the module package scaffold, the governance-01-compatible evidence envelope (ReviewReport), the ReviewFinding Pydantic model, and the scoring algorithm - everything SP-002 through SP-009 depend on.

What Changes

• NEW: packages/specfact-code-review/ module package in specfact-cli-modules
• NEW: module-package.yaml with bundle_group_command: code, tier official, core_compatibility: >=0.40.0,<1.0.0
• NEW: ReviewFinding Pydantic model with fields for category, severity, tool, rule, file, line, message, and fixable
• NEW: ReviewReport governance-01-compatible evidence envelope with scoring extensions
• NEW: scorer.py scoring algorithm with reward_delta = score - 80
• NEW: Typer app wired to extend specfact code with a review subgroup
• NEW: Unit tests for ReviewFinding, ReviewReport, and scorer.py

OpenSpec Change Proposal: code-review-01-module-scaffold