Serp Analysis

SERP analysis

You read the current top results for one keyword and report what it takes to win the page. You return an editorial angle and a brief, not a list of words to sprinkle in. The point is to find where the leaders are weak, not to copy them.

Initial assessment

• If .agents/blog-context.md exists, read it first. It sets the niche,

audience, and target country and language, and the SERP differs by all of them. Honor those over any default.

• Confirm the target keyword. If the user gave a URL instead, work out the

keyword it is built to rank for and analyze against that. If neither the user nor the context fixes a country and language, ask before guessing, or note your assumption in the output.

• Check what you can fetch. If a search or SERP tool is connected (for example an

MCP that returns results, or the ability to fetch result URLs directly), use it. If nothing can reach the web, say so plainly and ask the user to paste the top few competing pages so you can still run the analysis.

Method

1. Pull the top three to five organic results. Skip ads and skip pure

aggregators unless they genuinely compete. Note who ranks: the kind of site, not just the URL, since a forum thread and a vendor guide signal very different things about the page.

1. Read the dominant intent. From the titles and the pages themselves, decide

what the searcher actually wants: a how-to, a comparison, a definition, a buying decision, a tool. Get this wrong and depth will not save the piece.

1. Map the table stakes. List the subtopics and headings that nearly every

result covers. These are the sections a reader expects; leaving one out reads as a gap, not a choice.

1. Spot the optional angles. Note the subtopics only some pages cover. These are

where the field disagrees about what matters, and where a sharper take can live.

1. Find the gap. This is the work. Look for what is missing, thin, outdated, or

generic across the results: a step nobody shows, a claim nobody backs, advice that aged out, a question left hanging. Look for where you can be more specific, more current, or more credible than what ranks now.

Use the gap, do not mirror the structure

Matching the leaders point for point earns you a page that is merely as good as what already ranks, which is rarely enough to overtake it. Cover the table stakes so you are in the race, then win on the gap. One genuine advantage beats adding a tenth heading nobody asked for: first-hand testing, real numbers, a clearer walkthrough, a more honest comparison.

Output

Return a short brief in plain prose and tight lists:

• The dominant search intent, in one line, with the page type it implies.
• The must-cover subtopics, the table stakes a contender cannot skip.
• The gap: what is missing, thin, or dated across the current results, and where

you can be more specific or credible.

• A recommended depth to be competitive: a rough word-count range and a

heading-count range, drawn from what ranks, not a fixed target.

• The single sharpest angle to win, stated as the editorial promise of the piece.

This feeds straight into the article outline and pairs with keyword-research. For tuning a draft you already have against the live results, hand off to seo-optimize.

Related skills

blog-context · keyword-research · article · seo-optimize · ai-seo · house-style

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: RightBlogger
• Source: RightBlogger/bloggingskills
• License: MIT
• Homepage: https://bloggingskills.com/

Install and usage instructions live in the source repository linked above.