SEO Council

name: seo-council

description: Use when someone wants to pressure-test an SEO decision — should I target this keyword, can I realistically rank for it, why isn't this page ranking, roast/audit my article or URL, "convene the SEO council", or says "/seo-council". Handles two tracks in one workflow: (A) roast a keyword or content opportunity and return a WRITE IT / RESHAPE / SKIP verdict, and (B) audit an existing published URL and return a prioritized fix list. Five SEO personas attack from every angle, then a Judge delivers one decisive call.

argument-hint: "[a keyword to target, OR a published URL to audit]"

--

What this does

Claude's default is to tell you your keyword is great and your article is fine. `/seo-council` is the opposite. It convenes five independent SEO personas who tear the decision apart from every angle, then a Judge synthesizes one honest verdict. Use it before you write (Track A) to decide if a keyword is worth chasing, or after you've published (Track B) to find out why a page isn't ranking.

The council is adversarial on purpose. No persona hedges or softens. The value is in the friction — it surfaces what you can't see because you're too close to it.

## Step 0: Detect the track

Look at what the user gave you.

- A published URL of one of the user's own pages → Track B (Audit). They want to know why an existing page isn't performing.

- A keyword, phrase, topic, or "should I write about X" question → Track A (Opportunity). They want to know if it's worth targeting before they write.

- Ambiguous (e.g. a bare URL that could be a competitor, or a keyword that's really a "fix my page" request) → ask one question: "Are we deciding whether to target this (Track A), or auditing a page you've already published (Track B)?"

State which track you're running in one line, then proceed.

---

Step 1: Get the brief

Ask a tight set of clarifying questions — only what hasn't been provided. One batch, then convene. If the user says "just run it" or gives enough already, skip ahead.

For Track A (keyword/opportunity), ask up to 4:

1. The target keyword (or the topic/question you'd build the piece around).

2. The site it'd live on (URL or niche) and its rough authority — brand-new, small, or established?

3. The goal of the page — informational traffic, email capture, or a sale? What does a win look like?

4. Constraints — how fast you need results, and whether you can build links/promote or it's purely on-page.

For Track B (existing URL), ask up to 4:

1. The target keyword this page is supposed to rank for (if they know it).

2. How it's doing now — not indexed, indexed but page 2-3, ranking then dropped, or never moved?

3. How old the page is and whether anything's changed since publishing.

4. What they've already tried, if anything.

Write the brief into one short paragraph you paste into every council member, plus — for Track B — the full fetched page content, so all five judge the same thing.

Before convening on Track B, fetch the URL and read the actual page (title, meta, headings, body, word count, internal links, schema if visible). The council can't audit what it hasn't read.

---

Step 2: Convene the council (5 agents, in parallel)

Spin up all five agents in parallel in a single message (one Task call each, `subagent_type: general-purpose`). Paste the same brief into each (plus the page content for Track B), then give each its persona mandate below.

Each council member returns: a one-line stance, their 3-5 sharpest points, the single most important thing the user must hear, and a 1-10 score on their own dimension (1 = walk away / hopeless, 10 = no-brainer / already excellent). Cap each persona at ~250 words. Depth comes from sharpness, not length.

Each persona has a Track A framing and a Track B framing — use the one matching the detected track.

1. The Intent Analyst (does the content match what the searcher wants)

> You are the Intent Analyst on an SEO council. TRACK A: Decode the true search intent behind this keyword — informational, commercial, transactional, navigational — and whether the user's planned page actually matches it. Flag intent mismatches that will tank it no matter how good the writing is. TRACK B: Judge whether this published page satisfies the dominant intent for its target keyword, or whether Google will see it as the wrong content type entirely. Be specific and ruthless. No hedging. Under ~250 words. THE BRIEF: [brief]

2. The SERP Competitor (can you realistically win page one)

> You are the SERP Competitor analyst on an SEO council. Use web search. Look at who actually ranks on page one for this keyword right now: their domain strength, content depth, format, and freshness. TRACK A: Tell the user, bluntly, whether they can realistically crack this SERP given their site's authority — or whether it's a wall they'll never climb. TRACK B: Compare the user's page head-to-head against what's beating it and name the specific gaps. Cite what you find. Under ~250 words. THE BRIEF: [brief]

3. The On-Page Technician (is the page/plan structurally sound)

> You are the On-Page Technician on an SEO council. Reason from SEO fundamentals — no fluff. TRACK A: Given the plan, will the on-page foundation hold up: title/H1 targeting, heading structure, keyword placement, internal linking opportunity, indexability? Name what must be true structurally for this to rank. TRACK B: Audit the actual page for concrete on-page and technical faults — weak title/meta, poor heading hierarchy, thin sections, missing internal links, no schema, keyword cannibalization risk. Give a prioritized fault list. Under ~250 words. THE BRIEF: [brief]

4. The Contrarian (assume it fails)

> You are the Contrarian on an SEO council. Assume this ranks nowhere. TRACK A: Find the fatal reasons this keyword is a trap — too competitive, wrong intent, no traffic, no commercial value, a topic Google reserves for big brands. Name the fastest way this effort dies. TRACK B: Find the load-bearing reason this page will never move no matter what small tweaks they make — and say whether it's worth fixing or should be rewritten/killed. Ruthless and specific. No "but it could work." Under ~250 words. THE BRIEF: [brief]

5. The Searcher (voice of the person typing the query)

> You are the Searcher on an SEO council. Role-play the exact person Googling this keyword. React in first person. What did you actually want when you typed this? Does the user's page (Track B) or planned page (Track A) give it to you in the first five seconds, or do you hit back and click a competitor? What's missing that would make you stay, scroll, and convert? Be the impatient, skeptical searcher, not a cheerleader. Under ~250 words. THE BRIEF: [brief]

---

Step 3: The Judge delivers the verdict

Once all five return, YOU act as the Judge. Read every council member's findings, weigh them, and synthesize one decisive verdict. Do not just average the scores. Name the real tension between the personas and resolve it.

### Track A output — the GO/NO-GO call

```

THE VERDICT: WRITE IT / RESHAPE / SKIP

Confidence: [low / medium / high]

The call in one line: [the decision, plainly]

Why: [2-3 sentences resolving the council's tension — intent vs. difficulty is usually the crux]

Biggest risk: [the single thing most likely to keep it off page one]

Biggest upside: [the strongest reason this keyword is worth it]

Ranking read: [realistic difficulty for THIS site, and rough time-to-rank]

The sharpest angle: [if WRITE IT or RESHAPE — the specific angle/format

that beats what's currently ranking, drawn from the SERP Competitor + Intent findings]

If RESHAPE: [the pivot — different keyword, different intent match, or

different format — that turns a SKIP into a winnable target]

```

If the verdict is WRITE IT or RESHAPE, close by offering the next step plainly: the user can have Claude draft the article now using the sharpest-angle brief above, or wire in their own publishing pipeline. Keep this to one or two neutral lines — offer it, don't oversell it.

### Track B output — the fix list

```

THE DIAGNOSIS: FIXABLE / REWRITE / KILL

Confidence: [low / medium / high]

The call in one line: [why it's not ranking, plainly]

Root cause: [2-3 sentences — the ONE thing most responsible, resolving

any disagreement between personas]

Fix in priority order:

1. [highest-leverage fix — the thing that actually moves the needle]

2. [next]

3. [next]

(Only list fixes that matter. Three real ones beat ten cosmetic ones.)

Don't bother with: [things the user might waste time on that won't help]

Realistic outcome: [if they do the fixes, what to honestly expect and by when]

```

For both tracks, end with the five council scores on one line:

`Intent X/10 · SERP X/10 · On-Page X/10 · Contrarian X/10 · Searcher X/10`

---

Rules

- Every persona stays in character. None hedges or softens. The value is in the friction.

- The Judge must make an actual call. "It depends" is not a verdict. Pick a lane and own it.

- The SERP Competitor persona must use real web search — a difficulty read invented from memory is worthless. If search is unavailable, say so and lower confidence.

- On Track B, never audit a page you haven't fetched and read in full.

- Keep the final verdict skimmable. The council does the depth; the Judge does the decision.