CataloguePerformance, Scale & Cost
Frontend Performance & Bundle Audit
Reads the frontend as a user receives it: the bundle, the images, and the work the browser should not be doing.
What it examines
- What the production bundle actually contains
- Duplicate and unused dependencies shipped to the browser
- Scripts and styles that block first paint
- Images without a size or a modern format
- Largest paint, interaction delay, and layout shift
- One JavaScript file for every route
- Work done in the browser that belongs on the server
What you get back
A findings file. Each item has the evidence, how serious it is, and a fix.
Dependencies
- An AI coding agent that reads SKILL.md files, such as Claude Code, Codex or Gemini CLI
- Read access to the repository
- Optional: a production build, meaning minified bundles, source maps or a bundle analyser report. Without it, the bundle-size and duplicate-package checks are recorded as insufficient evidence.
- Optional: real user monitoring data, such as CrUX or a RUM export. A Lighthouse or WebPageTest report is accepted as lab data and labelled that way.
- Python 3 with the jsonschema package, to check findings.json before fixing anything
How it works
The audit is read-only. It reads the frontend code for heavy bundles, unoptimized images, render-blocking assets and missing code splitting.
With a production build it checks what is really in the bundle. With real user or lab data it compares LCP, INP and CLS with the standard thresholds and says which kind of data each number came from.
You get findings.json and findings.md. The remediate skill reads findings.json. Simple fixes such as pruning and images come first, and you approve each change.
Setup
- Unzip the download. Copy the audit and remediate folders into your agent's skills folder (for Claude Code, .claude/skills/) in the repository you want to check.
- Run your production build and keep the output folder, if you want the bundle checks. Export real user data or a Lighthouse report if you have one.
- Open the repository in your agent and ask it to run the c5-frontend-performance-audit skill.
- Check the result with the validator: pip install jsonschema, then python3 schema/validate.py findings.json.
- Ask your agent to run the c5-frontend-performance-remediate skill using findings.json. Fix some findings and leave the rest open if you like.
Run it with your agent
Copy the prompt, fill in the parts in square brackets, and paste it into your agent.
Run the c5-frontend-performance-audit skill on this repository. Production build output: [path, or none]. Real user monitoring data or a Lighthouse report: [path, or none]. Mode: audit only. Do not change any code. Write findings.json and findings.md in the repository root.
Run the c5-frontend-performance-remediate skill using findings.json. Work through the findings I choose, and leave the rest open.