github-code-review

---

LIBRARY-FIRST PROTOCOL (MANDATORY)

Before writing ANY code, you MUST check:

Step 1: Library Catalog

• Location: .claude/library/catalog.json
• If match >70%: REUSE or ADAPT

Step 2: Patterns Guide

• Location: .claude/docs/inventories/LIBRARY-PATTERNS-GUIDE.md
• If pattern exists: FOLLOW documented approach

Step 3: Existing Projects

• Location: D:\Projects\*
• If found: EXTRACT and adapt

Decision Matrix

Match	Action
Library >90%	REUSE directly
Library 70-90%	ADAPT minimally
Pattern exists	FOLLOW pattern
In project	EXTRACT
No match	BUILD (add to library after)

---

STANDARD OPERATING PROCEDURE

Purpose

Provide a GitHub-native code review that cites file:line evidence, maps to standards, and outputs actionable comments or summaries suitable for PR discussion threads.

Trigger Conditions

• Positive: reviewing GitHub PRs, generating review summaries, or preparing merge-blocking feedback.
• Negative: single-file lint fixes or runtime debugging (route to functionality-audit).

Guardrails

• Confidence ceiling: Include Confidence: X.XX (ceiling: TYPE Y.YY) with ceilings {inference/report 0.70, research 0.85, observation/definition 0.95}.
• Evidence discipline: Each comment must include file:line, severity (blocker/major/minor/nit), and a reference (style guide, security rule, performance budget).
• Structure-first: Keep examples/ and tests/ aligned with GitHub review formats (inline comments and summaries).
• Adversarial validation: Re-check for false positives; mark ambiguous items as “needs maintainer confirmation.”

Execution Phases

1. Scope & Context
   • Identify PR intent, risk areas, and excluded paths (generated/vendor).
   • Determine which findings are merge-blocking vs. advisory.
2. Review Passes
   • Pass 1: High-risk scan (security, data integrity, migrations).
   • Pass 2: Correctness and test adequacy.
   • Pass 3: Performance, maintainability, readability.
3. Evidence & Comment Drafting
   • For each finding, capture file:line, evidence type, severity, and fix guidance.
   • Group related issues to reduce comment noise.
4. Summary & Confidence
   • Provide an overall recommendation (approve/request-changes/comment) with rationale.
   • State residual risks and confidence with explicit ceiling.

Output Format

• Per-finding entries with severity, file:line, evidence type, and references.
• Consolidated PR summary with blockers and follow-ups.
• Confidence statement using ceiling syntax.

Validation Checklist

• Scope, exclusions, and merge criteria documented.
• Findings include file:line, severity, and references.
• Blockers vs. advisories separated; noise minimized.
• Confidence ceiling provided; output in English.

Confidence: 0.72 (ceiling: inference 0.70) - SOP rewritten with Prompt Architect confidence discipline and Skill Forge structure-first review steps.