skills.rest
jrc1883
jrc1883COMMUNITY
Explore all skills by this author →
GitHub

cross-model-review

Pairs Claude Code and Codex to generate normalized advisory reports for PR workflows.

Updated Aug 23, 2026
Search Tags:#automation#git#code-review#cross-model#advisory#outside-voice
One-click install
npx skills add https://github.com/jrc1883/popkit-ai --skill cross-model-review
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cross-model-review
Source: https://github.com/jrc1883/popkit-ai/tree/main/packages/popkit-ops/skills/pop-cross-model-review
Command: npx skills add https://github.com/jrc1883/popkit-ai --skill cross-model-review

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires popkit_shared, and includes scripts (resource) components.

What problem does it solve?

Cross-model advisory reviews provide an outside-voice assessment by using the opposite model family when possible, enriching PR feedback and reducing blind spots.

Core Features & Use Cases

  • Cross-model advisory reviews pair Claude Code with Codex (and vice versa) to generate a normalized advisory report and a readable advisory artifact.
  • Reviews can be printed to the console and optionally published as PR comments, with artifacts stored for traceability keyed to the current head SHA.
  • Use cases include pre-PR readiness checks, mid-flow outside-voice reviews during branch finishing, and on-demand requests for alternative-model feedback.

Quick Start

Run the cross-model review workflow from the project root with the default provider settings.

Frequently Asked Questions about cross-model-review

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

FAQPage Schema
What is a cross-model advisory review for pull requests?▼

Cross-model code review works by pairing the active model family with the opposite one, such as running Codex against Claude Code, to generate a normalized advisory report that reduces blind spots.

How do I run a cross-model code review in my PR workflow?▼

Run the cross-model review workflow from the project root with default provider settings to generate a normalized advisory report and optionally publish the results as PR comments.

Does cross-model review support both Claude Code and Codex?▼

Yes, cross-model review supports both Claude Code and Codex, pairing the opposite model families to provide outside-voice assessments during PR workflows and branch finishing.

Can I publish cross-model review results as PR comments automatically?▼

Yes, you can configure publication behavior via CLI parameters to optionally publish advisory results as PR comments and store artifacts for head SHA-based tracking.

When should I use an outside-voice model assessment for branch finishing?▼

Use an outside-voice model assessment during branch finishing for mid-flow reviews and pre-PR readiness checks, utilizing artifact normalization to ensure traceability keyed to the current head SHA.

Related Skills

Explore More Software Engineering Skills→
evolution-foundationevolution-foundation
520

dev-ask

Query one target model for a labeled advisory answer.

Official
Intermediate
nanomadernanomader
1

pr-codex-review-poll

Poll head-scoped Codex review findings for the current PR head.

Community
Intermediate
qq824045252-ship-itqq824045252-ship-it

superpowers-finishing-a-development-branch

Integrate completed feature branches via merge, PR, keep, or discard workflows.

Community
Advanced
neuron-oneneuron-one
25

adversarial-review

Route code changes to a different AI model for independent review.

Community
Advanced