Post

Agent-run snippets solve the old problem with console pastes: nobody reran them after a deploy. An LCP check on every preview build catches the hero image regression before users do.

Totally agree, and you can use the CLI version webperf-snippets.net/CLI FYI, I have it on the roadmap to publish a GitHub Action

A GitHub Action posting LCP and CLS as a PR comment would put the budget in front of the reviewer before merge, which is when a regression is cheapest to fix.

That shared contract (status, value, rating, issues) is why this is a major version: if you read the output from code, some fields have changed. 56 snippets, 7 of them new: compression, Early Hints, Speculation Rules, third-party impact by domain, and more.

For agents: • A "Copy prompt" button next to the snippets, so an agent runs them with the CLI and reports back • Agent Skills for Claude Code, Gemini CLI and other agents • A Markdown version of every page, llms.txt and WebMCP tools #WebMCP #SKILLS

The CLI reaches 1.0: audit, loading and interaction workflows, CrUX data, and a GitHub Action with LCP and CLS budgets. `npx webperf-snippets example.com` webperf-snippets.net/CLI #CLI