grill

Performs automated adversarial code reviews on pull requests.

2|Updated May 10, 2026
One-click install
npx skills add https://github.com/loomantix/codex-platform --skill grill-loomantix
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: grill
Source: https://github.com/loomantix/codex-platform/tree/main/.codex/skills/grill
Command: npx skills add https://github.com/loomantix/codex-platform --skill grill-loomantix

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill addresses the risk of shipping bugs, security vulnerabilities, and architectural regressions by automating a rigorous, multi-lane adversarial review process for draft pull requests.

Core Features & Use Cases

  • Adversarial Review Lanes: Executes specialized review passes including silent failure detection, security analysis, and tenant-coupling checks.
  • Inline Finding Ledger: Automatically posts verified findings directly to the PR diff, tracks fixes, and manages resolutions.
  • Flexible Review Depth: Offers a Lean mode for standard PRs and a Deep mode for high-risk changes involving complex logic or sensitive data.

Quick Start

Run the grill skill on your current open draft pull request to initiate an adversarial review and generate an inline finding ledger.

Frequently Asked Questions about grill

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

FAQPage Schema
What is adversarial code review for pull requests?

Adversarial code review executes specialized passes like silent failure detection and tenant-coupling verification on draft pull requests to hunt bugs, security flaws, and architectural inconsistencies.

How do I automate security analysis on a GitHub pull request?

You can automate security analysis on a GitHub pull request by running an adversarial review skill that posts verified findings directly to the PR diff and manages resolutions through an inline finding ledger.

Can I check for tenant-coupling issues in my PR before merging?

Yes, tenant-coupling verification is a specialized lane in the adversarial review process that checks your draft pull requests for architectural inconsistencies and coupling issues before integration.

Does this adversarial PR review work with standard and high-risk code changes?

Yes, the review supports flexible depth modes: a Lean mode for standard pull requests and a Deep mode for high-risk changes involving complex logic or sensitive data.

What is the best way to track and resolve code review findings on a GitHub PR?

The best way to track findings is to use an inline finding ledger that automatically posts verified issues to the PR diff, tracks fixes, and manages resolutions directly within the GitHub workflow.