seo-content-production（SEO 内容 production）

SEO Content Production

Complete workflow from keyword research to publication. This skill orchestrates other skills - it does not duplicate their functionality.

Workflow Overview

RESEARCH → STRUCTURE → SOURCES → DRAFT → QUALITY → PUBLISH
    │           │          │        │        │         │
seomachine   outline   proprietary  write   quality   webflow
                       content              -loop     -publish

Phase 1: Research

Use seomachine for Data

All SEO data comes from the seomachine skill. DO NOT hardcode credentials.

Credentials: Stored in vault .env file:

• DATAFORSEO_LOGIN, DATAFORSEO_PASSWORD - Keyword research
• GA4_PROPERTY_ID, GOOGLE_SERVICE_ACCOUNT_PATH - Traffic data
• GSC_SITE_URL - Search Console data

Common research tasks:

# Keyword research
import sys
sys.path.insert(0, ".claude/skills/seomachine/modules")
from dataforseo import DataForSEO

dfs = DataForSEO()
ideas = dfs.get_keyword_ideas("waldorf vs montessori", limit=50)
questions = dfs.get_questions("homeschool curriculum", limit=20)
serp = dfs.get_serp_data("best homeschool math curriculum")

# Generate content brief with competitor analysis
python3 .claude/skills/seomachine/scripts/content_brief_generator.py "keyword phrase"

# Find keyword gaps vs competitors
python3 .claude/skills/seomachine/scripts/competitor_gap_finder.py --batch --min-volume 200

Validation Criteria

Before proceeding:

• Search volume > 500/month
• Keyword difficulty < 30 (winnable)
• Topic aligns with OpenEd expertise
• We have (or can create) proprietary perspective

Phase 2: Structure

Identify OpenEd's Unique Angle

Before outlining, answer:

1. What does OpenEd believe about this topic? (Check opened-identity skill)
2. What can we say that competitors can't? (Podcast quotes, staff expertise)
3. What do most articles get wrong or avoid?

Content Type Templates

Templates are in references/content-templates.md. Common types:

Phase 3: Source Compilation

Search Order (Priority)

1. OpenEd Book - .claude/references/OpenEdBook/
2. Published Content - Search Master Content Index
3. Podcast Transcripts - Grep Published Content/Podcasts/
4. Staff Knowledge - Tag relevant team members
5. External Research - Fill remaining gaps

Source Search Commands

# Search published content
grep -rli "topic" "Published Content/" | head -20

# Check Master Content Index
grep -B5 "Tags:.*homeschool" ".claude/references/Master_Content_Index.md"

Phase 4: Draft

Before Writing

• Research brief complete (from seomachine)
• Unique angle identified
• Sources compiled
• Internal links planned (3+ minimum)

Writing Constraints

Apply these skills automatically:

• ai-tells - Patterns to avoid
• ghostwriter - Authentic voice techniques

Key rules:

• BLUF first - lead with the answer, not the backstory (see references/bluf-rewrite-pattern.md)
• Table early (for skimmers)
• 5+ internal links, 2-3 external
• Use spaced hyphens - not em dashes

AI-isms to Eliminate

See references/ai-tells-quick-reference.md for full list.

Kill words: delve, comprehensive, crucial, leverage, landscape, navigate, foster, facilitate, realm, paradigm, journey, tapestry, myriad, seamless

Kill patterns:

• "It's not just X - it's Y" (correlative constructions)
• "In today's fast-paced world..."
• "Let's dive in..."
• "The best part?..."

Phase 5: Quality Control

Use quality-loop Skill

Run every draft through 5 judges:

Quick Checks (Grep Patterns)

# Forbidden words
grep -inE "delve|comprehensive|crucial|leverage|landscape" DRAFT*.md

# Correlative constructions
grep -inE "isn't just|not just .* - it's" DRAFT*.md

# Em dashes (should be hyphens with spaces)
grep -n "—" DRAFT*.md

# Internal link count
grep -oE '\(https://opened\.co[^)]*\)|\(/[^)]*\)' DRAFT*.md | wc -l

Phase 6: Visuals

Use nano-banana-image-generator Skill

Comparison article concepts:

• Two hands offering different gifts
• Split path through two environments
• Same tree with different root systems

Phase 7: Publish

Use webflow-publish Skill

Pre-publish checklist:

• All blocking judges passed
• Thumbnail created
• Meta elements finalized
• Internal/external links verified

Post-Publish

1. Update articles that should link TO this piece
2. Sync Master Content Index
3. Schedule social promotion via newsletter-to-social

Skill Dependencies

Project Folder Structure

Studio/SEO Content Production/[Topic]/
├── SEO_RESEARCH_BRIEF.md   # From seomachine
├── SOURCES.md              # Compiled proprietary sources
├── OUTLINE.md              # Approved structure
├── DRAFT_v1.md             # First draft
├── DRAFT_v2.md             # Post-quality-loop
└── thumbnail-final.png     # Header image

Quick Start

1. Validate topic with seomachine
   → Volume > 500, KD < 30, OpenEd has perspective

2. Create folder: Studio/SEO Content Production/[Topic]/

3. Run research: seomachine content brief + competitor gap

4. Compile sources: OpenEd book → Published → Podcasts → External

5. Get outline approved: Structure + unique angle

6. Draft with constraints: ai-tells + ghostwriter + internal links

7. Quality loop: 5 judges, fix blocking issues

8. Create thumbnail: nano-banana-image-generator

9. Publish: webflow-publish → newsletter-to-social

*Refactored: 2026-01-29 (consolidated from seo-content-writer, removed credential exposure)*