ck:code-review

Identify security holes and failure modes in code changes via adversarial reasoning.

Updated Aug 23, 2026

Search Tags:#security #quality #review #diff #code-review #adversarial

Ready to run

npx skills add https://github.com/dungnt97/wallet-portal --skill ck-code-review-dungnt97

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ck:code-review
Source: https://github.com/dungnt97/wallet-portal/tree/main/claude/skills/code-review
Command: npx skills add https://github.com/dungnt97/wallet-portal --skill ck-code-review-dungnt97

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Code reviews often miss subtle flaws and security risks; this skill applies adversarial reasoning to identify weaknesses, false assumptions, and failure modes, with an always-on red-team perspective.

Core Features & Use Cases

Adversarial, evidence-based critique that actively tries to break the code
Supports multiple input modes: PR diffs, commit diffs, pending changes, and full codebase scans
Structured findings with severity, impact, and remediation guidance to strengthen the codebase

Quick Start

Provide a target (PR, commit hash, or pending changes) and run the skill to generate an adversarial review.

Frequently Asked Questions about ck:code-review

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

FAQPage Schema

How do I run an adversarial code review on a PR diff?▼

Run an adversarial code review on a PR diff by providing the pull request as the target. The review actively tries to break the code to identify and document security holes, logical flaws, and failure modes.

What is adversarial code review and how does it find security holes?▼

Adversarial code review applies an always-on red-team perspective to identify weaknesses, false assumptions, and failure modes. It actively tries to break the code to find subtle security risks and logical flaws that standard reviews often miss.

Can I scan a full codebase for security flaws and logical errors?▼

Yes, you can scan a full codebase for security flaws and logical errors. The review supports multiple input modes including PR diffs, commit hashes, pending changes, and full codebase scans to guide targeted reviews.

How do I get remediation guidance for code vulnerabilities found in pending changes?▼

To get remediation guidance for code vulnerabilities, provide your pending changes as the target. The review returns structured actionable findings with clear severity, impact, and specific remediation guidance to strengthen the codebase.

Does this code review approach work with commit diffs?▼

Yes, this code review approach works directly with commit diffs. You can provide a commit hash as the target to generate an evidence-based critique that documents failure modes and security holes in those specific changes.

What is the best way to find false assumptions in code changes?▼

The best way to find false assumptions in code changes is using an adversarial, evidence-based critique. This approach actively attempts to break the code, returning structured findings with severity and impact to strengthen the codebase.