claude-review-runtime

Preflight MCP availability and delegate code, documentation, and configuration changes to the claude-cross-review-protocol.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/nowonbun/nowonbun-harness --skill claude-review-runtime
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-review-runtime
Source: https://github.com/nowonbun/nowonbun-harness/tree/main/codex-skills/runtime-management_claude-review-runtime
Command: npx skills add https://github.com/nowonbun/nowonbun-harness --skill claude-review-runtime

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps agents decide when Claude MCP cross-review is required and enforces local preflight checks before handing code, documentation, configuration, or governance changes to the Claude review protocol, preventing invalid or oversized requests and blocking reviews when prerequisites are missing.

Core Features & Use Cases

  • Applicability Classification: Determine whether an operation requires plan-review, source-review, result-review, or re-review and enforce required ordering (e.g., plans-review before execution).
  • Local Preflight Checks: Verify MCP availability and responsiveness (expecting an OK response) and block with clear reasons if unavailable or unresponsive.
  • Input Preparation Rules: Enforce concise request bodies, UTF-8 integrity checks for text files, path-based references instead of full file dumps, and review goal composition per phase.
  • Protocol Delegation & Reporting: Delegate execution to the claude-cross-review-protocol skill for normalization and logging, and produce user-facing summaries including phase, cross-review status, summary, key findings, and next action.
  • Use Case: Run a plans-review to validate an implementation plan before execution, perform a source-review prior to finalizing a change, or run a result-review after applying file modifications.

Quick Start

Use the claude-review-runtime to preflight MCP availability, classify the review as source-review, prepare concise UTF-8-safe targets, and hand off execution to the claude-cross-review-protocol.

Frequently Asked Questions about claude-review-runtime

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is a preflight check for Claude MCP code review?

A preflight check for Claude MCP code review validates local MCP availability and responsiveness before delegating changes, ensuring the review protocol receives concise, UTF-8-safe inputs and blocks execution if prerequisites are missing.

How do I classify a code review phase for a Claude cross-review?

You classify a code review phase by determining whether the operation requires plan-review, source-review, result-review, or re-review, enforcing required ordering such as completing plans-review before execution and source-review before finalizing changes.

How do I prepare inputs for a Claude MCP cross-review protocol?

You prepare inputs for a Claude MCP cross-review by enforcing concise request bodies, verifying UTF-8 integrity for text files, using path-based references instead of full file dumps, and composing review goals specific to the classified phase.

Can I run a Claude cross-review if the local MCP is unresponsive?

You cannot run a Claude cross-review if the local MCP is unresponsive because preflight checks expect an OK response to proceed, blocking the delegation with clear reasons when availability and responsiveness validation fails.

What reporting fields does a Claude cross-review runtime return?

A Claude cross-review runtime returns user-facing summaries that include the review phase, cross-review status, summary, key findings, and next action after delegating execution to the claude-cross-review-protocol for normalization and logging.

When do I need a result-review versus a source-review in a cross-review workflow?

You need a source-review prior to finalizing a change to validate code, documents, or configuration, whereas a result-review is performed after applying file modifications to assess the outcomes of the applied changes.