Audits and improves organic search performance — technical health, site architecture, internal linking, structured data, and the content decisions that determine what can rank. Use this to run an SEO audit, diagnose why pages are not ranking or were deindexed, plan a site's URL and navigation structure, add structured data, or prioritize which SEO fixes are worth doing.
npx skills add https://github.com/cbrock84/headcount --skill seo-strategy
Skipping to content when the problem is technical wastes months. Work down; stop when you find the
break.
crawl budget wasted on parameter and filter URLs.
pointing at the wrong page and unintended duplicates are the two usual causes.
and still lose because it answers a different question than the query asks.
Targeting them is a decision to spend without return.
Structure follows how people look for things, not how the company is organized.
duplicate content in the crawler's eyes.
Internal linking is the most underused lever available: it is fully within your control, and it
directs authority to the pages you actually care about. Link with descriptive anchor text from your
strongest pages to the ones that need help.
Speed and mobile usability are ranking inputs and, more importantly, conversion inputs. Audit them
as part of every SEO review rather than treating them as an engineering concern.
Measure on field data from real users where available, not only lab tools. Lab scores on a fast
connection routinely miss what mobile users experience.
The three that matter:
and format, render-blocking resources, and slow server response.
fixed by reducing main-thread JavaScript work.
space for images, ads, and embeds before they load.
Mobile is the version that gets indexed. Check that mobile serves the same content as desktop —
content hidden or dropped on mobile is content that does not exist for ranking — and that tap
targets, font sizes, and viewport configuration do not make the page unusable.
Prioritize performance work by pages with traffic, not by worst score. A terrible score on a page
nobody visits is not a finding.
Mark up what genuinely exists on the page — the entity, its attributes, the relationships. It
clarifies meaning to the crawler and can earn richer results.
Use JSON-LD in the page head rather than inline microdata — it is easier to maintain and validate,
and it does not entangle markup with content.
The types worth implementing, in rough order of return: organization and site identity, breadcrumbs,
articles, products with offers and availability, frequently asked questions, events, local business
details, and job postings. Implement the properties each type actually requires plus the recommended
ones you can populate honestly; partial markup with missing required fields is usually ignored
entirely.
Validate after every template change. Structured data breaks silently when a template is edited, and
nothing tells you.
Never mark up content that is not visible on the page, and never mark up ratings, prices, or events
that are not real. Both are policy violations with penalties attached.
Rank fixes by traffic at risk against effort. In practice the order is almost always: indexation
problems, then intent mismatch on high-value pages, then internal linking, then structured data,
then everything else.
Findings by severity with the evidence for each, the sequenced fix list, expected impact and when it
should be visible, and what you are choosing not to fix.
Take cbrock84/seo-strategy from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.