0 min read

Generate a CRO Audit

Designers

Claude Skill

TL;DR:

I built this as a Claude skill that audits a live page against a six-part conversion framework, traces every claim to what the page actually does, not what it appears to promise, and produces a scored, prioritized HTML report.

Why This, Specifically

Checking whether a page's claims actually hold up, whether a discount banner's link leads to a discount, whether a testimonial appears anywhere, is tedious to do thoroughly by hand and easy to skip. This skill runs a fixed six-part framework, traces every element back to what the page actually does, and cites the benchmark it's comparing against. The judgment calls, whether an offer reads as compelling, whether credibility actually lands, still need your read.

How It Works

Six-part framework: credibility, lure, objection handling, social proof, ease of use, result. Each checked against its own definition, not a fixed list of examples.
Branches by business model: e-commerce, lead-gen, SaaS trial, subscription, marketplace, or high-ticket B2B. What counts as a strong offer differs by model.
Traces claims to their actual mechanic: a CTA gets checked against what it links to, not what it appears to promise.
Technical depth scales with what's available: qualitative signals by default, or real Core Web Vitals and live-rendered checks when a browser tool is connected.
Optional competitor and benchmark passes: compare against a named competitor or let the skill find and confirm comparable ones, and freshness-check the cited benchmark against a live source.
PIE-scored recommendations, each linked back to the finding it fixes.

The Skill

Setup interview, live fetch and verification of the target, the six-element assessment, any optional passes approved, then a single HTML report with findings and fixes cross-linked, screenshots included when a live browser check ran.

When and How To Use It

Works on a single page or a multi-step flow (cart through checkout, a full visitor-to-buyer path). Built for Claude Code.

Call it directly by name, e.g.:

use the cro-audit-generator skill on [page or flow]

A Note on What This Skill Can and Cannot Do

No connection to Google Analytics, Clarity, or any analytics platform. Findings only reflect real visitor behavior if you supply that data yourself. The technical section stays qualitative without a connected browser tool, the report says which mode it ran in. Check the report against the actual page before acting on it.

This skill is a structured prompt. It instructs Claude to evaluate without softening criticism, but it does not guarantee accuracy. Claude can still miss things, misread context, or produce confident-sounding findings that turn out to be wrong. The output is a starting point for your own judgment, not a substitute for it.
Treat any finding that would drive a real decision as something to verify independently before acting on it.

Get the Skill

The skill is available free on GitHub.

Download from GitHub

Let's Talk