A senior-level, security-first AI code review agent that audits codebases, pull requests, and implementation diffs following strict xAI standards.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install grok-code-review
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install grok-code-review using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Grok Code Review?
Grok Code Review is an advanced static analysis and auditing tool designed to run inside AI coding environments. By acting as a senior+ security researcher and systems architect, this skill automatically reviews pull requests, code snippets, or full files for critical vulnerabilities, performance bottlenecks, and style inconsistencies. It is one of the premier Openclaw Skills for developers looking to inject rigorous engineering standards directly into their local or CI/CD coding workflows.
The skill goes beyond simple syntax checking; it provides copy-pasteable fixed versions for identified issues and clearly separates must-fix blockers from nice-to-have cleanups. This ensures that every automated audit provides actionable, context-aware instructions to elevate the overall quality and security posture of your software projects.
Grok Code Review Use Cases
- Pre-Commit Code Audits: Automatically run a rigorous security check on local modifications before committing to the repository.
- Pull Request Evaluation: Integrate with development workflows to assess code diffs, identifying security flaws, race conditions, or hardcoded secrets.
- Refactoring Validation: Verify that large-scale code restructurings did not introduce new vulnerabilities or performance bottlenecks.
- Onboarding Codebase Sweeps: Gain deep insights into legacy systems or external repositories by executing a comprehensive static analysis.
How Grok Code Review Works
- Context Gathering: The agent reads the target code block, file, or diff using native environment tools like read_file, grep, or list_dir to establish complete context.
- Security & Quality Checklist Evaluation: It systematically evaluates the code against a rigorous 10-point checklist covering injection vectors, concurrency, resource leaks, authentication, and dependency safety.
- Structured Report Generation: The agent formats the findings into a clear, standardized document featuring executive summaries, severity-categorized issues, and positive commendations.
- Actionable Resolution: For every problem discovered, the tool provides specific code blocks showing the optimal fix, alongside a definitive workflow recommendation (e.g., Approve, Request Changes).
Grok Code Review Setup
To integrate this agent into your workflow, ensure your Openclaw Skills environment is configured correctly. Define the skill's capabilities in your local system configuration:
# Example command to register the grok-code-review skill with your local environment
openclaw register-skill grok-code-review
Ensure that environment tools like read_file and grep are accessible to allow the agent to fetch multi-file context during deep scans.
Grok Code Review Data Schema & Taxonomy
The grok-code-review skill utilizes a structured system metadata definition and outputs human-readable reports. Here is how its data taxonomy is organized:
| Property | Type | Description |
|---|---|---|
name |
String | Unique identifier: grok-code-review |
version |
String | Semantic versioning value (e.g., 0.1.0) |
emoji |
String | Graphic tag: 🔎 |
scope |
String | static-analysis-only (no arbitrary execution allowed) |
Report Output Schema
Each audit report follows a strict Markdown structure containing:
- Summary: High-level review overview.
- Critical / High-Severity Issues: Severe threats (RCE, leaks) with fixed code block templates.
- Medium Issues: Non-critical bugs or performance problems.
- Low / Polish / Style: Linter, styling, and design notes.
- What Was Done Well: Commendations for solid architectural choices.
- Recommendation: Strict final status (
Approve,Approve with minor comments,Request changes, orMajor rework required).
Grok Code Review Advanced Features
- Deep-Context Crawling: Uses filesystem utilities dynamically to fetch referenced helper functions or interface declarations outside the current viewport.
- Copy-Paste Code Injection: Provides fully formed, secure code snippets directly within the issue descriptions for instant fixing.
- xAI Standard Verification: Audits codebases utilizing specialized security rules inspired by cutting-edge AI security frameworks.
- ClawHub-Safe Guardrails: Built-in static-only mode guarantees the agent will never execute unverified commands or run untrusted build processes.
SKILL.md
Loading
Related Openclaw Skills
Claw Test Skill
A streamlined, side-effect-free validation skill engineered to test and confirm direct API publishing workflows on the ClawHub registry.
Conversation Video Generator
An automated workflow tool designed to synthesize multi-speaker conversational videos from structured transcript manifests, featuring synchronized voice audio and timed text overlays.
Go Like Hell
An executive-level AI agent skill that translates the legendary Ford-Ferrari motorsport rivalry into actionable business frameworks for handling competition, culture clashes, and high-pressure decision-making.
The Code Breaker
An executable AI knowledge companion that guides you through the science of CRISPR gene editing, patent competition, research dynamics, and the ethics of altering human DNA.
The Road to Unfreedom: Disinformation & Geopolitics AI Skill
An AI-powered critical thinking and information-hygiene framework designed to detect disinformation, analyze authoritarian ideologies, and defend public truth.
Chat Record Generator
An automated multi-agent simulation tool that generates realistic WeChat group conversation histories in Excel format for model training and platform testing.
Featured*
