dcarrero/awesome-code-review-prompts

This repository is a collection of ready to paste prompts for asking an AI assistant to review your code. Instead of writing your own instructions each time, you copy one of these prompt files into a chat with an AI model and paste your code or diff alongside it. The prompts are organized by dimension: security, cybersecurity and infrastructure, performance, accessibility, usability, code quality, testing, and safety concerns specific to AI generated code. There is also an all in one prompt that runs every dimension at once and produces a single scored report. On top of these general prompts, the project includes add ons for over a dozen languages and frameworks, such as JavaScript, Python, Ruby, PHP, Java, Go, C sharp, Swift, Kotlin, and Rust, which you append after a general prompt to get checks specific to that stack. Every prompt file exists in both English and a matching Spanish version, so teams working in either language can use the same structure. The suggested workflow is simple: pick a dimension, paste the master prompt into the chat, add your code as a git diff for a quick pull request check or a full file or folder for a deeper audit, then optionally stack a language specific add on. Each master prompt is written to first ask the AI up to three clarifying questions before it reviews, so the review is based on real context rather than guesses. The prompts are plain text and not tied to one AI vendor. The README states they are tuned for Claude Fable 5 but are also designed to work with other frontier models that can follow structured instructions, including other Claude models, GPT models, Gemini, Grok, and DeepSeek. The project is licensed under MIT.