corvus-review-r2

Orchestrate multi-pass PR reviews across architecture, correctness, security, and conventions.

3|Updated Dec 18, 2025
One-click install
npx skills add https://github.com/NachoFLizaur/corvus --skill corvus-review-r2
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: corvus-review-r2
Source: https://github.com/NachoFLizaur/corvus/tree/main/skill/corvus-review-r2
Command: npx skills add https://github.com/NachoFLizaur/corvus --skill corvus-review-r2

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill coordinates a comprehensive PR review workflow, aligning architecture, correctness, security, and conventions into a repeatable, multi-pass process that yields structured findings.

Core Features & Use Cases

  • Orchestrates four review passes: architecture, correctness, security, and conventions, with a shared context across passes.
  • Supports parallel execution for passes 1-3 and sequential pass 4 after completion, ensuring timely, high-quality feedback.
  • Produces standardized REVIEW_FINDINGS to facilitate traceability, auditing, and gating of PRs across teams.
  • Adapts to large or complex PRs requiring cross-cut analysis and consistent quality gates across codebases.

Quick Start

Run the Corvus review with a PR context to produce R2 findings.

Frequently Asked Questions about corvus-review-r2

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

FAQPage Schema
How do I automate multi-pass PR reviews for large codebases?

Multi-pass PR reviews coordinate parallel evaluations across architecture, correctness, security, and conventions, followed by a sequential consolidated polish pass to produce structured findings for large PRs.

What is a multi-pass code review workflow and when do I need it?

A multi-pass code review workflow evaluates a PR sequentially or in parallel across distinct dimensions like architecture and security, needed for large collaborative software projects requiring cross-cut analysis and consistent quality gates.

How do I structure code review findings for better traceability and PR gating?

Structuring code review findings involves generating standardized REVIEW_FINDINGS outputs after evaluating PR passes, facilitating traceability, auditing, and gating across teams by providing clear, consolidated feedback.

Can I run architecture, security, and correctness PR reviews in parallel?

Yes, parallel PR review execution allows passes 1 through 3 covering architecture, correctness, and security to run simultaneously, with a final pass 4 polishing step executing sequentially after their completion to ensure timely feedback.

Does a structured multi-pass review workflow work for complex collaborative PRs?

A structured multi-pass review workflow adapts to large or complex collaborative PRs by applying cross-cut analysis, shared context across evaluation passes, and consistent quality gates across codebases to yield high-quality feedback.

What are the limitations of multi-pass PR review orchestration?

Multi-pass PR review orchestration requires shared context across passes to function effectively, and its multi-phase sequencing may introduce overhead for small PRs that do not require cross-cut architectural or security analysis.