skills.rest
weiping
weipingCOMMUNITY
Explore all skills by this author →
GitHub

cross-modal-review

Route work products to a different reviewing model for Contract-based quality checks.

1|Updated May 9, 2026
Search Tags:#quality gate#AI code review#diff review#contract validation#adversarial security#cross-modal review#refusal routing
One-click install
npx skills add https://github.com/weiping/gbrain-cn --skill cross-modal-review-weiping
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cross-modal-review
Source: https://github.com/weiping/gbrain-cn/tree/main/skills/cross-modal-review
Command: npx skills add https://github.com/weiping/gbrain-cn --skill cross-modal-review-weiping

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It reduces the risk of shipping incorrect, unsafe, or Contract-inconsistent work by having a different AI model review the output before it’s finalized.

Core Features & Use Cases

  • Contract-based judging: The reviewer grades work against the originating skill’s Contract section rather than subjective impressions.
  • Refusal-safe multi-model routing: If one model refuses, the skill silently switches to the next model in the chain.
  • Code-review support with adversarial mode: For diffs, it can hand off to Codex review or run an adversarial challenge to look for security and injection issues.
  • User sovereignty enforcement: Reviewer findings are informational only; the user explicitly approves what to apply.

Quick Start

Ask the agent to run a cross-modal review of the change against the originating Contract, including security-focused adversarial checks for risky diffs.

Frequently Asked Questions about cross-modal-review

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

FAQPage Schema
How do I get an independent AI code review before committing changes?▼

A cross-modal review validates changes by routing them to a different reviewing model for an independent second opinion. It grades code against a Contract and runs adversarial security checks.

What is Contract-based judging in AI diff review?▼

Contract-based judging grades work against the originating skill's Contract section rather than subjective impressions. It enforces structured, evidence-based findings for code review validation.

How do I handle model refusal when running automated security checks on code?▼

Refusal-safe multi-model routing silently switches to the next model in the chain if one refuses. This ensures adversarial security checks complete without interrupting the review workflow.

Does AI code review automatically apply fixes to my codebase?▼

No, AI code review enforces user sovereignty by keeping reviewer findings informational only. You must explicitly approve what suggestions to apply, ensuring manual control over the codebase.

When should I use a second-model quality gate for my code?▼

Use a second-model quality gate for significant code changes, security-sensitive edits, skill creation, repeated churn, and brain-page quality checks. Apply it prior to bulk operations to prevent shipping unsafe work.

Related Skills

Explore More Software Engineering Skills→
ngochuy13ngochuy13

cross-modal-review

Review work products against an originating skill's Contract with PASS/ISSUES grading.

Community
Intermediate
Tx1207Tx1207
5

~evolve

Generate skill and preference candidates from project changes and evidence.

Community
Advanced
bhianerdbhianerd
1

codex

Review code diffs with OpenAI Codex and flag critical issues.

Community
Advanced
jefferyjobjefferyjob
31

openclaw-it-team-deploy

Deploy the openclaw-it-team multi-agent workflow with localized openclaw.json and Feishu credentials.

Community
Intermediate