The toolkit

Every worksheet, checklist and script used to build the case study in the book. Free to download, no account required.

agent

• AI Agent Brief: Drop this in your repo and point Claude Code, Cursor, Aider or Copilot at it. Thirteen numbered rules with exact limits, the required schema per page type, the suggested working order, and an explicit list of things the agent must never do, starting with inventing a rating.
• Agent Prompts: Nine copy-paste prompts: head-tag audit, schema audit and repair, static-rendering check, depth audit, snippet rewrites, building a new page from the spec, reading your GSC export, internal-link audit, and a pre-deploy gate for CI.
• Machine-Readable Rules: The same rules as structured data, so an agent checks them deterministically rather than by vibes. Includes the position/CTR diagnostic as a decision table and the measurement warnings that stop it misreading your dashboard.

analysis

• Search Console Analyser: Point it at your own Search Console export and it runs every diagnostic in this book on your data: traffic concentration, brand vs non-brand split, wasted visibility, per-page verdict of on-page / authority / snippet, and the one thing to do next. No dependencies, nothing uploaded, runs on your machine. Add --json to hand the output to an agent.

default

• Page Spec Worksheet: One copy per page. All twelve fields in the order you should write them, each with its formula and a checklist. The title comes last because it is a compression of the direct answer.
• Static Page Generator: Working, dependency-free build script. Turns a list of page definitions into complete static HTML with head tags, body copy and JSON-LD, plus a sitemap and robots.txt. One data structure produces both the visible content and the markup, so they can never disagree. Includes the lint pass that catches over-long titles and missing direct answers.
• Schema Snippets: Copy-paste JSON-LD for WebApplication, Product, BreadcrumbList, HowTo, FAQPage, Organization, WebSite and BlogPosting, with the AggregateRating guard clause and the rules that keep you out of trouble.
• Query Research Sheet: Sort by takeable, not by volume. Columns for the weakest page-one result and its quality score, because that is the slot you are actually competing for.
• Five-Day Diagnostic: Run this before building anything. Demand, supply, decomposition, differentiator, instrumentation, with the gates that tell you to stop.
• Launch Asset Pack: Exact image dimensions and all four copy lengths. Half of directory rejections are dimension problems this prevents. Includes the Product Hunt and Show HN prep list.
• Directory Checklist: The twelve worth a day, the comparison directories worth an hour, and the badge tier worth one weekend and never a second. Plus what not to touch.
• Outreach Tracker: Target, contact, why this target, sent, follow-up due, outcome. Follow up once after five to seven days, then stop.
• Disavow Template: Grouped and commented, including a prominent warning that most sites should never submit one. Prepare it, keep it, do nothing unless a manual action appears.
• Pre-Deploy Checklist: Forty seconds before every deploy, plus monthly and quarterly checks. Ends with the four mistakes that actually cost this project, including the one that is not technical and is the most expensive on the list.