review

Orchestrates a parallel code-review swarm and generates a unified fix/suggestion report.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/CasaPerks/casaflow --skill review-casaperks
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review
Source: https://github.com/CasaPerks/casaflow/tree/main/core/skills/review
Command: npx skills add https://github.com/CasaPerks/casaflow --skill review-casaperks

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill reduces the risk of shipping broken code by running a structured, parallel code review that escalates correctness issues into a single actionable report.

Core Features & Use Cases

  • Parallel specialist swarm review: Dispatches multiple review agents simultaneously, each focused on a concern using configured models and file globs.
  • Diff scoping for speed: Filters review work to only changed files/hunks and can run a fast-pass tier for per-task gates.
  • Deep logic verification (pre-PR): Runs a skeptical correctness reviewer after the swarm to find bugs that pattern-based reviewers miss.

Quick Start

Run the review as part of your workflow by invoking: /review.

Frequently Asked Questions about review

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

FAQPage Schema
How do I automate code review to catch correctness bugs before a PR merge?

Code review automation orchestrates a parallel specialist swarm to analyze filtered git diffs, running a deep logic verification stage to catch correctness bugs before a PR merge and generating a unified summary report.

Can I scope code review to only changed files and hunks for faster per-task quality gates?

Yes, code review can be scoped to only changed files and hunks by generating filtered diffs from git changes, enabling a fast-pass tier designed specifically for per-task quality gates before PR submission.

How does parallel specialist code review work with configured file globs?

Parallel specialist code review dispatches multiple review agents simultaneously, with each agent focusing on a specific concern by applying configured models and file globs to isolate relevant code sections for analysis.

What is the best way to deduplicate code review findings by file and line?

The best way to deduplicate code review findings by file and line is through a structured review swarm that escalates correctness issues into a single actionable report, skipping or running stages based on configured tiers.

Does deep logic verification find bugs that pattern-based code reviewers miss?

Deep logic verification runs a skeptical correctness reviewer after the parallel swarm stage to find bugs that pattern-based code reviewers miss, specifically targeting pre-PR code correctness before merges.

How do I integrate Jira-native workflow quality gates into my pre-PR code review?

Jira-native workflow quality gates integrate into pre-PR code review by applying mechanical scoring and reporting requirements, selecting specialists by configured tiers and globs to produce a unified summary with fix or suggestion sections.