Can I use this for specific audits like security or performance?

Yes, the --focus flag allows you to narrow the review specifically to security, performance, architecture, or edge-case categories.

What is the difference between Harsh and Constructive tones?

Constructive tone provides supportive, educational feedback that 'cares personally,' while Harsh tone offers brutally honest, sarcastic critique to highlight issues directly.

How does Candid Review know my project's specific coding standards?

It automatically looks for a Technical.md file in your project root or .candid directory to load and enforce your team's specific rules and patterns.

Does it support automatic commits for suggested fixes?

Yes, when enabled via CLI flags or configuration, the skill can automatically commit applied fixes with a detailed summary and co-author tags.

Does it work with existing Git branches?

Absolutely. It can detect staged changes, unstaged changes, or compare your current branch against main, develop, or other configured merge targets.

Candid Code Review

Name: Candid Code Review
Author: ron-myers

byron-myers

•

Collaboration Tools

Performs automated code reviews using the Radical Candor framework to provide actionable, tone-aware feedback based on project standards.

Candid Code Review acts as an AI-powered full-stack architect that evaluates your code changes before they reach production. By implementing the 'Radical Candor' philosophy—care personally while challenging directly—this skill identifies architectural flaws, security risks, and performance bottlenecks. It integrates seamlessly with your Git workflow, automatically detecting staged or unstaged changes and comparing them against your project's unique requirements defined in Technical.md. Whether you need a supportive guide or a brutally honest critique, Candid helps you maintain high-quality codebases with targeted focus modes and automated fix applications.

Key Features

0114 GitHub stars

02Deep Git integration for staged, unstaged, and branch-level diffs

03Specialized focus modes for security, performance, and architecture

04Optional auto-commit functionality for applied fixes with detailed logs

05Configurable review tones ranging from constructive to 'senior dev' harsh

06Automated enforcement of project-specific standards via Technical.md

Use Cases

01Onboarding new developers by enforcing team-specific architectural patterns

02Conducting pre-commit audits to catch bugs and standards violations early

03Automating the remediation of common code smells and performance bottlenecks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ron-myers/candid candid-review

For use in Claude.ai and ChatGPT

Download Skill