Claude Code prompt tip: 100/100 Lighthouse score with 2 extra words
Adding just two words to a prompt helped Claude Code achieve a perfect score on website performance tests. This shows how small tweaks to instructions can make AI coding agents much more effective.
A developer discovered that Claude Code initially missed some optimizations, scoring around 91 or 92 on Lighthouse tests. By adding two specific words to the prompt, the AI was able to identify missing steps and reach a perfect 100 score. This highlights that even advanced AI models sometimes need a little nudge to perform their best. Other developers are now testing if similar small prompt changes work with tools like Codex CLI to solve complex coding tasks.
Key points
- Small prompt changes can lead to big performance gains in AI-generated code.
- Claude Code achieved a perfect 100 Lighthouse score after a tiny instruction tweak.
- AI agents sometimes miss steps unless specifically told to focus on them.
- This technique is being explored for other developer tools like Codex CLI.
Quick term guide
- AI coding agents
- AI tools that can help write, edit, or organize software code.
- AI coding agent
- An AI tool that can write, edit, and run code from your instructions.
- coding agents
- AI programs designed to autonomously perform tasks like writing or fixing code.
- coding agent
- An AI tool that writes or edits code from a person’s instructions.
- Lighthouse
- A tool that tests how well a website performs and follows best practices.
- developers
- Developers are people who build software, apps, or websites.
- Codex CLI
- A command-line tool for using OpenAI Codex from a terminal.
- AI agents
- AI agents are AI tools that can carry out steps toward a goal, not just answer once.