Loading...
The playbook Moonrank's SEO Agent uses to run SEO as a continuous loop — find opportunities across every layer, prioritize by impact, ship the fixes, measure what moved, and start again.
Download the methodology, see how the agent runs the weekly cycle, or run it live inside the SEO Agent.
The SEO skill is Moonrank's master operating loop: how the SEO Agent turns technical audits, quick wins, content, and measurement into one weekly cycle — a single ranked list of work, shipped and measured, instead of four backlogs nobody sequences.
It's written as a SKILL.md, a plain-language instruction format that people and AI agents can both follow. Read it to run the loop yourself, use it to structure how your team works, or let the SEO Agent execute the full cycle end to end.
Downloading the playbook is free with a Moonrank account. Running the loop every week without anyone having to remember — that's exactly what the SEO Agent is for.
“We've done three audits in two years and traffic hasn't moved. What are we doing wrong?”
The audits found the problems; nobody ran the loop. Of the 74 issues from the last one, 3 are worth shipping this cycle: a canonical fix on the category template covering 600 URLs, titles on 8 posts ranking 6–9, and one decayed guide down 55% year over year. Shipped and measured over two-week cycles, that's an estimated 3,100 extra clicks a month by cycle four — and a scored list waiting for cycle five.
Search Console at minimum; conversion data if you want the loop to chase revenue, not just clicks.
The agent records the before, then pulls opportunities from every layer into one list.
Everything scored by traffic × confidence ÷ effort; the top items get shipped and logged.
Two weeks later, each change is judged against its baseline and the results reweight the next cycle.
Clicks, impressions, and positions per page group, so every change has a before.
Technical issues, quick wins, decaying pages, and content gaps in a single list.
Everything scored by expected traffic × confidence ÷ effort — one list, not four.
Every change recorded with a date, so results stay attributable.
Each shipped change labelled won, lost, or unclear against its own baseline.
Fix types that pay off on your site get more weight in the next cycle's scoring.