Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Turn a target keyword into a fill-in brief a writer can execute: search intent, target and per-section word counts, the H2/H3 outline, entities and questions to cover, internal and external link suggestions, and a title tag plus meta description set. Use whenever the user says "write a content brief", "content brief for X", "SEO brief", "outline for this keyword", "what should this article cover", or is about to commission or write a page and wants it built to outrank the current top 10. Reach f
.claude/skills/whatsuppiyush-seo-content-brief/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 29% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 50% | 0% |
Give it a target keyword and it studies what ranks, then writes a brief a writer can execute: per-section word counts, entities and questions to cover, and a title/meta set.
Claude Code
---
name: seo-content-brief
description: >-
Turn a target keyword into a fill-in brief a writer can execute: search intent, target and
per-section word counts, the H2/H3 outline, entities and questions to cover, internal and
external link suggestions, and a title tag plus meta description set. Use whenever the user
says "write a content brief", "content brief for X", "SEO brief", "outline for this
keyword", "what should this article cover", or is about to commission or write a page and
wants it built to outrank the current top 10. Reach for it before drafting any SEO article
or landing page. Output is a ready-to-hand-off brief template.
---
# SEO Content Brief
Study what already ranks for a keyword, then write a brief that tells a writer exactly what to
produce to beat it. The brief is derived from evidence (the current top 10 results), not from
imagination: word counts come from what ranks, the outline comes from the coverage patterns
across competitors plus the gaps, and the questions come from what searchers actually ask.
## When to use this
- About to write or commission an SEO article, guide, or landing page.
- A writer needs direction beyond "write about X".
- Trigger phrases: "content brief", "SEO brief", "outline for", "what should this cover",
"brief for this keyword".
## Workflow
### 1. Read the top 10 ranking pages
Pull the current top 10 organic results for the target keyword and study them, not just skim.
For each, note: page type (guide, listicle, product, comparison), rough word count, the H2/H3
headings, and what it covers well or badly. This is the bar you have to clear.
### 2. Determine search intent
From the shape of the top 10, name the dominant intent so the page matches it. If the SERP is
how-to guides, an informational deep-dive wins; if it is listicles, a ranked list wins; if it
is product pages, a commercial page wins. A well-written page that fights the SERP's intent
does not rank. State the intent and the winning page type explicitly.
### 3. Set target and per-section word counts
Take the median word count of the relevant top-10 pages (ignore outliers and off-intent
pages) and aim slightly above it, only if the extra words add substance. Longer is not better;
matching intent thoroughly is. Then split the target across sections so the writer knows the
weight of each part (intro, each H2 block, conclusion). This prevents a bloated intro and a
starved core.
### 4. Build the H2/H3 outline
Merge the heading structures of the top competitors into a superset, then add what they all
miss. Every H2 should map to a sub-intent or question a searcher has. Order them the way a
reader wants the information, usually: what/why up top, how in the middle, edge cases and FAQ
toward the end. Mark which headings are table-stakes (everyone covers them) versus your
differentiation (the gap you fill).
### 5. List entities, questions, and subtopics to cover
Give the writer the concrete things the page must mention to read as authoritative:
- **Entities:** the specific tools, concepts, people, standards, and terms that ranking pages
reference. Search engines expect a real article on this topic to name these.
- **Questions:** the People Also Ask questions for the keyword plus related questions from the
SERP. Each is a candidate H2 or FAQ item.
- **Subtopics and gaps:** what the top 10 under-cover or get wrong; this is your edge.
### 6. Suggest internal and external links
- **Internal:** 3 to 6 existing pages on the site this article should link to, with suggested
anchor text, to build topical connection and pass equity.
- **External:** 2 to 4 authoritative, non-competitor sources to cite (studies, docs,
primary data). Citing real sources raises trust and is easy for the writer to skip without
a prompt.
### 7. Write the title tag and meta description
Provide a title tag (aim ~55 to 60 characters, keyword near the front, a reason to click) and
a meta description (~150 to 155 characters, keyword plus the benefit and an implied CTA).
Offer one primary and one alternate of each.
## OUTPUT TEMPLATE
Primary keyword: <keyword> Secondary keywords: <kw>, <kw>, <kw> Search intent: <informational / commercial / transactional> - winning page type: <type> Target word count: <N> (competitor median <M>)
### <H3> ### <H3>
...
Internal:
External (cite):
Title tag (A): <~57 chars> Title tag (B): <alternate> Meta description (A): <~155 chars, keyword + benefit + implied CTA> Meta description (B): <alternate>
## Worked example
Target keyword: **how to do a backlink audit**.
The top 10 are all how-to guides (intent = informational, page type = step-by-step guide),
median ~1,900 words. Brief sets target at 2,100 words. The merged outline yields H2s: what a
backlink audit is, when to run one, the step-by-step process, tools you can use, what to do
with toxic links, how often to repeat it, plus an FAQ. The differentiation gap: every
competitor lists tools but none shows a decision rule for disavow-versus-ignore, so that
becomes a flagged differentiation H2.
Must-cover entities: referring domains, anchor text distribution, disavow file, DR/DA, toxic
links, Google Search Console. PAA questions ("Is a backlink audit worth it?", "How do you
find bad backlinks?") become FAQ items. Internal links point to the site's disavow guide and
its anchor-text article; external cites go to primary Search Console documentation.
Title tag: "How to Do a Backlink Audit (Step-by-Step, 2026)". Meta: "Run a backlink audit in
7 steps: find toxic links, read your anchor profile, and decide what to disavow. A repeatable
process, no guesswork."From God of Skills: a curated, hand-tested directory of AI skills, prompts, templates and image style guides. Source: https://godofskills.com/skills/seo-content-brief?ref=claude-skill
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 22,056 | 12,223 | -45% | 1 | 1 | 0% | 4,216 | 4,150 | -2% | 0 | 0 | — |
case-01 | fail→pass | 19,805 | 14,420 | -27% | 1 | 1 | 0% | 3,700 | 4,587 | +24% | 0 | 0 | — |
case-03 | fail→pass | 20,799 | 15,810 | -24% | 1 | 1 | 0% | 3,962 | 5,094 | +29% | 0 | 0 | — |
case-04 | fail→fail | 35,129 | 35,617 | +1% | 1 | 1 | 0% | 6,184 | 8,013 | +30% | 0 | 0 | — |
case-05 | fail→fail | 9,714 | 6,392 | -34% | 1 | 1 | 0% | 1,721 | 2,933 | +70% | 0 | 0 | — |
case-06 | pass→pass | 12,815 | 11,933 | -7% | 1 | 1 | 0% | 2,148 | 3,656 | +70% | 0 | 0 | — |
case-22 | pass→pass | 17,481 | 10,965 | -37% | 1 | 1 | 0% | 2,863 | 3,547 | +24% | 0 | 0 | — |
case-07 | fail→pass | 17,162 | 12,583 | -27% | 1 | 1 | 0% | 3,039 | 4,038 | +33% | 0 | 0 | — |
case-08 | fail→pass | 21,593 | 11,854 | -45% | 1 | 1 | 0% | 3,706 | 4,062 | +10% | 0 | 0 | — |
case-09 | fail→pass | 15,605 | 12,893 | -17% | 1 | 1 | 0% | 2,706 | 4,067 | +50% | 0 | 0 | — |
case-10 | fail→pass | 16,166 | 12,027 | -26% | 1 | 1 | 0% | 2,856 | 3,827 | +34% | 0 | 0 | — |
case-11 | pass→pass | 18,453 | 12,805 | -31% | 1 | 1 | 0% | 3,106 | 3,933 | +27% | 0 | 0 | — |
case-12 | pass→pass | 14,786 | 14,710 | -1% | 1 | 1 | 0% | 2,581 | 4,223 | +64% | 0 | 0 | — |
case-13 | fail→pass | 15,618 | 11,390 | -27% | 1 | 1 | 0% | 2,699 | 3,712 | +38% | 0 | 0 | — |
case-14 | fail→pass | 20,593 | 18,360 | -11% | 1 | 1 | 0% | 3,154 | 4,886 | +55% | 0 | 0 | — |
case-15 | fail→pass | 16,471 | 13,772 | -16% | 1 | 1 | 0% | 2,863 | 4,306 | +50% | 0 | 0 | — |
case-16 | fail→pass | 15,189 | 16,919 | +11% | 1 | 1 | 0% | 2,498 | 4,750 | +90% | 0 | 0 | — |
case-17 | pass→pass | 15,311 | 11,527 | -25% | 1 | 1 | 0% | 2,437 | 3,695 | +52% | 0 | 0 | — |
case-18 | fail→pass | 17,634 | 12,846 | -27% | 1 | 1 | 0% | 2,872 | 3,917 | +36% | 0 | 0 | — |
case-19 | fail→pass | 20,873 | 12,691 | -39% | 1 | 1 | 0% | 3,088 | 3,688 | +19% | 0 | 0 | — |
case-20 | fail→pass | 17,026 | 16,704 | -2% | 1 | 1 | 0% | 2,628 | 4,356 | +66% | 0 | 0 | — |
case-21 | fail→pass | 22,995 | 18,919 | -18% | 1 | 1 | 0% | 3,732 | 5,057 | +36% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +64 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/6/2026 | +23% |
Other measured skills in the registry, with their headline benchmark lift.