What does an on-page audit cover?

Eight dimensions, each scored Pass, Needs work, or Fail with a specific fix: the title tag, the meta description, header structure, body content, internal links, images and media, the URL slug, and on-page schema. The audit reads the rendered HTML rather than just the visual page, because the served markup is what crawlers actually see, and it works from one target query and the page's role (commercial, informational, or navigational).

What makes a good title tag and meta description?

A title tag is one unique title per URL, roughly 50 to 60 characters before SERPs truncate it, with the primary query near the front, the brand at the end if it earns inclusion, and distinct from (not word-for-word identical to) the H1. A meta description is one unique description per URL, roughly 150 to 160 characters, treated as ad copy that earns the click with a soft CTA where natural, restating the value proposition rather than repeating the title.

How should headers be structured?

Exactly one H1 per page that contains or paraphrases the primary query, H2 sections covering the major sub-topics, and H3 and below used only when an H2 has genuine sub-points. Do not skip levels (no H2 followed directly by an H4), and write headers descriptive enough that a reader could navigate the page by them alone. The header outline is both a relevance signal and a usability one.

Does adding the keyword more times help?

No. Keyword density is not a real ranking signal, so stuffing the target query five more times into the body does nothing useful. Prioritize relevance and topical depth instead: answer the primary user intent in the first paragraph, cover the topic at least as comprehensively as the pages that currently rank, include related entities and supporting concepts naturally, and write for the reader rather than the crawler.

What should I check before optimizing a page?

That the page is crawlable and indexable, because no on-page work helps a noindexed page; that it serves a real user intent, because a page answering no genuine query will not rank however well-tagged it is; and that no other page on the site targets the same query, which is cannibalization. If the target query itself is unknown, run keyword research or ask for one before optimizing, since on-page work without a query is guesswork.

Skill · On-page SEO

On-page SEO.

Optimize one page for relevance, click-through, and crawler comprehension.

Optimize a single page for search relevance, click-through, and crawler comprehension. A complete audit covers eight dimensions, each scored Pass, Needs work, or Fail with the specific fix noted: title tag, meta description, header structure, body content, internal links, images and media, URL slug, and on-page schema.

The work starts from one target query and reads the served markup, not the visual page alone. It assumes the page can already be crawled and indexed; on-page optimization cannot rescue a page search engines will not index.

Audience: SEO and content teams auditing or optimizing a single page, rewriting titles and metas, or improving a page's click-through from search.

View the skill on GitHub Browse the full catalog

The framework

Eight dimensions, scored with a specific fix.

A complete on-page audit covers all eight. Score each Pass, Needs work, or Fail, and note the exact change.

01Title tag: one unique title per URL, roughly 50 to 60 characters, the primary query near the front, brand at the end if it earns inclusion, and distinct from the H1.
02Meta description: one unique description per URL, roughly 150 to 160 characters, written as ad copy that earns the click with a soft CTA, restating the value rather than the title.
03Header structure: exactly one H1 that contains or paraphrases the query, H2s for major sub-topics, no skipped levels, and headers a reader could navigate by alone.
04Body content: the primary intent answered in the first paragraph, depth that matches or exceeds the competition, related entities included naturally, and no keyword stuffing.
05Internal links: 2 to 3 outbound and 2 to 3 inbound links to closely related pages, with descriptive anchor text, pointing at canonical URLs rather than redirects.
06Images and media: descriptive alt text and file names, a modern format, lazy loading below the fold, and width and height set to prevent layout shift.
07URL slug: lowercase and hyphen-separated, the query included naturally, short, no dates unless time-bound, and no tracking parameters or random hashes.
08On-page schema: the appropriate Schema.org type with required properties filled, validated in the Rich Results Test, and matching what is visible on the page.

The discipline

Relevance over density, indexability first.

On-page work assumes the page can actually be found. Optimizing a page that is noindexed or uncrawlable is wasted effort, so confirm the technical layer first: no title or header fix helps a page search engines will not index.

Keyword density is not a real ranking signal. Adding the target query five more times into the body does nothing, while relevance and topical depth do everything. Answer the primary intent in the first paragraph, cover the topic at least as deeply as what already ranks, and include related entities naturally rather than writing for the bot.

Two pages targeting the same query is cannibalization, and it splits the ranking signal between them. When that happens, decide which page should own the query rather than optimizing both against each other; that decision is a content-audit call, not an on-page one.

Reference files

Three references that go alongside the SKILL.md.

references/audit-template.md
A fillable on-page audit template, copy and use.
references/onpage-checklist.md
A quick-reference checklist for the eight dimensions.
references/title-and-meta-patterns.md
Patterns for writing strong title tags and meta descriptions.

Browse all reference files on GitHub

Bridges to other skills

The skills on either side of the page.

On-page work is single-page scope. These cover the layer beneath it, the query that defines it, and the multi-page decisions around it.

The layer beneath
seo-technical
Site-wide crawl, indexing, rendering, and speed sit there. Confirm a page is indexable in the technical layer before doing on-page work on it.
Defines the query
seo-keyword
Finds and classifies the target query this page is optimized for. If the query is unknown, run keyword research before optimizing.
Reads the SERP
seo-competitor
Analyzing the competitive SERP and where rivals win is a wider exercise. On-page sets the depth bar by matching what already ranks for one query.
Many pages at once
seo-content-audit
Auditing many pages for prune, merge, or cannibalization decisions is the audit's job. When two pages fight for one query, that skill decides which wins.

Open source under MIT

Read the SKILL.md on GitHub.

The skill source lives in the rampstackco/claude-skills repository alongside dozens of other skills covering the full lifecycle of brand and product work. This page is a structured overview; the SKILL.md is the source. MIT licensed.

View SKILL.md Browse the full catalog

Frequently asked questions.

What does an on-page audit cover?: Eight dimensions, each scored Pass, Needs work, or Fail with a specific fix: the title tag, the meta description, header structure, body content, internal links, images and media, the URL slug, and on-page schema. The audit reads the rendered HTML rather than just the visual page, because the served markup is what crawlers actually see, and it works from one target query and the page's role (commercial, informational, or navigational).
What makes a good title tag and meta description?: A title tag is one unique title per URL, roughly 50 to 60 characters before SERPs truncate it, with the primary query near the front, the brand at the end if it earns inclusion, and distinct from (not word-for-word identical to) the H1. A meta description is one unique description per URL, roughly 150 to 160 characters, treated as ad copy that earns the click with a soft CTA where natural, restating the value proposition rather than repeating the title.
How should headers be structured?: Exactly one H1 per page that contains or paraphrases the primary query, H2 sections covering the major sub-topics, and H3 and below used only when an H2 has genuine sub-points. Do not skip levels (no H2 followed directly by an H4), and write headers descriptive enough that a reader could navigate the page by them alone. The header outline is both a relevance signal and a usability one.
Does adding the keyword more times help?: No. Keyword density is not a real ranking signal, so stuffing the target query five more times into the body does nothing useful. Prioritize relevance and topical depth instead: answer the primary user intent in the first paragraph, cover the topic at least as comprehensively as the pages that currently rank, include related entities and supporting concepts naturally, and write for the reader rather than the crawler.
What should I check before optimizing a page?: That the page is crawlable and indexable, because no on-page work helps a noindexed page; that it serves a real user intent, because a page answering no genuine query will not rank however well-tagged it is; and that no other page on the site targets the same query, which is cannibalization. If the target query itself is unknown, run keyword research or ask for one before optimizing, since on-page work without a query is guesswork.