creo-seo-content — Next.js content quality specialist

Content-specific SEO auditor. Runs as standalone or as the Phase 2 delegate of creo-seo. Scores each page on 5 content dimensions, detects AI-generated patterns, maps search intent, scores opportunity, and produces rewrite priorities with before/after examples.

Commands

Command	Purpose
/creo seo-content audit <url>	Full content audit on a single URL or a list
/creo seo-content humanity <url>	Humanity score only (AI-pattern detection)
/creo seo-content intent <url or keyword>	Search intent classification
/creo seo-content linking <url>	Internal linking audit
/creo seo-content rewrite <url>	Generate rewrite priorities with before/after
/creo seo-content gate <mdx-path>	Publish gate — pass/fail score + reasons

Scoring

Composite 0–100 per page:

Content_Quality =
  (Humanity        × 0.30) +
  (Specificity     × 0.25) +
  (Structure       × 0.20) +
  (SEO_Compliance  × 0.15) +
  (Readability     × 0.10)

Publish threshold: ≥ 80 with zero Critical issues.

Audit flow

1. Load project profile from .claude/skills/creo-seo/creo-seo-{project_id}.md (shared with creo-seo skill). Fall back to fresh scan if missing.
2. Fetch page content — WebFetch for live URLs, Read for local MDX/TSX files.
3. Normalize text — strip HTML, extract main content, skip nav/footer.
4. Score 5 dimensions per references/content-quality-rubric.md (or shared ref in ../creo-seo/references/).
5. AI-pattern sweep per references/ai-pattern-detection.md — count phrases, vague words, boosters.
6. Structure analysis — H1/H2/H3 hierarchy, paragraph lengths, table/list presence.
7. Citability check per ../creo-seo/references/geo-citability.md — passages in 134–167 word range, 5-dimension scoring.
8. Intent classification — informational / navigational / transactional / commercial investigation.
9. Internal linking audit per ../creo-seo/references/internal-linking-recipe.md — count, anchors, cluster fit.
10. Output per-page report + rewrite priorities.

Search intent classification

4-class system:

Intent	Keyword signals	SERP features
Informational	what, why, how, guide, tutorial, tips, definition	featured snippet, knowledge graph, PAA, video, images
Navigational	login, sign in, website, dashboard, account, brand names	brand knowledge panel
Transactional	buy, purchase, pricing, free trial, coupon, order, download	shopping results, local pack, ads
Commercial investigation	best, review, vs, compare, alternative, similar, top	carousel, comparison tables, lists

Use intent to route:

• Informational → aim for citability + structured answers
• Commercial investigation → comparison tables, feature matrices
• Transactional → strong CTAs, trust signals, not just SEO
• Navigational → avoid SEO cannibalization of branded queries

Opportunity scoring (when GSC/keyword data available)

8-factor weighted score 0–100:

Opportunity =
  Volume(25%) + Position(20%) + Intent(20%) + Competition(15%)
  + Cluster(10%) + CTR(5%) + Freshness(5%) + Trend(5%)

Priority tiers:

• CRITICAL ≥ 80
• HIGH ≥ 65
• MEDIUM ≥ 45
• LOW ≥ 25
• SKIP < 25

Use with seo-google extension for live GSC data.

Writing patterns to enforce

Direct-answer opening (for AI citability)

Every H2 section must open with a 1–2 sentence direct answer before expanding. Pattern: "X is Y. It does Z."

Key Takeaways block (after intro, before first H2)

3–5 bullets of actual conclusions (not table of contents). AI summaries pull from here.

Mini-story injection

2–3 per long-form piece. Structure: named person + dates + outcome.

YouTube embed

≥ 1 relevant video per long-form article for multi-modal signal.

Content quality gates by page type

Page type	Min words	Target	Uniqueness
Homepage	500	800–1200	N/A
Service page	800	1200–1800	80%+
Blog post	1500	2000–2500	60%+
Product page	300	500–800	Per-SKU
Location page	500	600–800	60%+
Pricing	400	600–900	N/A

Reference library

Loaded on demand:

File	Purpose
references/search-intent.md	4-class intent system with signals
references/opportunity-scoring.md	8-factor content prioritization
references/content-playbook.md	Writing patterns (direct-answer, mini-stories, takeaways)
references/rewrite-templates.md	Before/after examples for common patterns
../creo-seo/references/content-quality-rubric.md	Shared 5-dim scoring
../creo-seo/references/ai-pattern-detection.md	Shared AI phrase/vague-word lists
../creo-seo/references/geo-citability.md	Shared citability rubric
../creo-seo/references/internal-linking-recipe.md	Shared linking rules

Report output

Save to {reports_path}/seo-content-{YYYYMMDD}-{HHMM}.md. Format:

# Content audit — {url}

**Composite score:** 68/100 (C)
**Publish status:** needs rework
**Intent:** commercial investigation (primary), informational (secondary)
**Opportunity score:** 72/100 (HIGH)

## Dimension breakdown
| Dimension | Score | Notes |
|-----------|------:|-------|
| Humanity | 58 | 12 AI phrases, 18 vague words in 1800w |
| Specificity | 62 | 2 stats only, no sources |
| Structure | 80 | Good H2s, paragraphs slightly long |
| SEO compliance | 75 | Keyword missing from first 100 words |
| Readability | 70 | Flesch 62, avg sentence 24 words |

## AI pattern flags (12 instances)
- "in today's digital world" — line 14
- "leverage" — line 28, 47
- ...

## Rewrite priorities (ranked by impact)
### 1. Opening paragraph — humanity + citability

BEFORE:
> {current opening}

AFTER:
> {rewritten opening with definition, stat, source}

### 2. Section 3 — specificity
...

## Internal linking gaps
- 2 of 3–5 links (below minimum)
- Missing pillar reference to /guide/podcast-hosting
- Orphan: 0 inbound links from other posts

## Next steps
1. Apply top 3 rewrites (est. score 68 → 82)
2. Add 1–2 internal links to pillar + related
3. Add 3 stats with citations to sections 2, 4, 6

Parallel execution

When called by creo-seo as part of /creo seo audit, run in parallel with Phase 3 + Phase 4. When called standalone, serial execution is fine.

Quality gates

• Every rewrite priority includes BEFORE + AFTER
• AI phrase flags are line-numbered
• Intent + opportunity scoring only fires when data is available; otherwise skip
• Reports saved to disk
• Scoring must be reproducible from refs (no magic constants in prose)