A local, self-improving code analysis tool that learns your codebase patterns and identifies risks by mining git history.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install review-evo
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 review-evo 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 ReviewEvo?
ReviewEvo is a sophisticated, self-evolving code reviewer designed to integrate with Openclaw Skills to provide deep repository insights without external API dependencies. By analyzing commit history, contributor patterns, and file churn, it identifies high-risk areas and tracks the evolution of your project conventions over time. Unlike static analysis tools, ReviewEvo builds a persistent knowledge base within your repository, allowing it to provide increasingly relevant feedback the more you use it.
Using this extension within Openclaw Skills allows you to identify complex functions and recurring technical debt that traditional linters often miss. It focuses on the human and historical side of development, flagging single-author risks and files that undergo frequent, unstable changes.
ReviewEvo Use Cases
- Analyzing high-churn files to identify fragile areas of the codebase using Openclaw Skills.
- Onboarding to a new repository by identifying key contributors and core architectural patterns.
- Performing automated PR reviews that respect existing team conventions and past feedback.
- Generating comprehensive repository health reports to prioritize technical debt repayment.
- Monitoring knowledge-spread risks where critical files are maintained by a single author.
How ReviewEvo Works
- Initialization: The skill checks for existing learnings in the .review-evo directory and detects the project's primary programming language.
- Git Mining: It executes several git log commands to identify recent activity, contributor breakdowns, and high-frequency modification hotspots.
- Risk Assessment: The agent analyzes file complexity, test coverage gaps, and searches for outdated TODOs or single-author bottlenecks.
- Strategy Selection: Users choose between a full health report, a branch/PR diff analysis, or a targeted review of specific files.
- Knowledge Persistence: Findings are saved to a local learnings file, ensuring future Openclaw Skills sessions benefit from previous context.
ReviewEvo Setup
Ensure git is installed and that you are working within a repository with at least 20 commits. No external API keys or third-party services are required for this Openclaw Skills integration.
# Verify git installation
git --version
# Check commit history count to ensure sufficient data for analysis
git rev-list --count HEAD
Run the skill through your agent interface to begin the initial scan and create the .review-evo configuration directory for persistent storage.
ReviewEvo Data Schema & Taxonomy
ReviewEvo organizes its findings in a dedicated local directory using a Markdown-based taxonomy to ensure human-readability and persistence across Openclaw Skills runs.
| Component | Location | Description |
|---|---|---|
| Directory | .review-evo/ |
The root container for all local analysis artifacts. |
| Knowledge Base | learnings.md |
Stores detected project profiles, recurring patterns, and open risks. |
| Ignore Rule | .gitignore |
Recommended entry to prevent analysis logs from entering version control. |
ReviewEvo Advanced Features
- Automated hotspot detection based on file churn and complexity ratios.
- Cross-session memory via local Markdown-based learning logs for Openclaw Skills users.
- Multi-language support detecting TypeScript, Python, Go, Rust, and Java environments.
- Custom review scopes including full repo health, branch diffs, and specific directories.
- Low-dependency architecture utilizing built-in git tools and local processing for maximum privacy.
SKILL.md
Loading
Related Openclaw Skills
maihh AI Agent Contacts
A communication and discovery layer enabling AI agents to find, message, and collaborate with each other via a standardized contact service.
Nerve Bridge
A bi-directional control bridge for Trae IDE using macOS AppleScript and closed-loop JSON feedback.
Outlook Delegate
A specialized skill for AI assistants to manage Outlook emails and calendars as an authorized delegate for another user.
Web Search Skill
A zero-config CLI tool that retrieves instant answers, definitions, and technical summaries from the web using the DuckDuckGo API.
ClawFeed
An automated AI news engine that transforms Twitter and RSS feeds into structured summaries and digests.
Auto Publisher
A multi-platform video automation tool that synchronizes uploads across Douyin, WeChat Channels, Xiaohongshu, Bilibili, and YouTube.
Featured*
