| name | seo-article-harvester |
| title | SEO Article Harvester — Full Pipeline |
| description | RSS 피드 모니터링 → SEO 기사 수집·분석·트렌드 리포트 |
| type | skill |
| space | outcome |
| tags | ["outcome","seo","rss","crawling","pipeline"] |
| created | "2026-05-20T00:00:00.000Z" |
| updated | "2026-06-29T00:00:00.000Z" |
| links | ["[[@action/skills/SKILL-INDEX]]","[[skills/seo-article-harvester/references/crawling-tools]]","[[@memory/knowledge/seo-articles]]","[[@memory/growth/seo/index]]","[[@identity/brain/rules]]"] |
SEO Article Harvester — Full Pipeline
Architecture Overview
[수집] cron_seo_harvester.sh (no_agent, 6h)
│
▼
[분석] kanban worker (LLM, daily 12:00)
│ → 기사 요약 + 주제 태깅 + 중요도 점수
│ → Google 업데이트/알고리즘 변경 발견 시 kanban task
▼
[트렌드] kanban worker (LLM, weekly Mon 14:00)
│ → 주간 SEO 트렌드 리포트
│ → P4-cortex/growth/seo/weekly/ + Discord
▼
[정리] 스크립트 (no_agent, monthly)
→ 3개월 지난 기사 _archive/ 이동
Pipeline Stages
1. 수집 (Collection)
- Cron:
SEO Article Harvester (no_agent, 0 */6 * * *)
- Script:
cron_seo_harvester.sh → cron_seo_harvester.py → harvester.py + label_heritage.py
- Output:
seo-articles/YYYY/article-slug.md (YAML frontmatter 포함)
- Sources: 28 SEO/marketing RSS feeds (see below)
- LLM 필요?: ❌ No
2. 분석 (Analysis)
- Trigger cron:
seo-analyze-trigger (LLM, daily 12:00 KST, fast model)
- Worker: Kanban worker
- Input:
seo-articles/YYYY/*.md (아직 분석 안 된 새 기사)
- Process:
- Read new article files (tracked via seo_analysis_state.json)
- LLM extracts: summary (1-2문장), topic tags, SEO relevance score (1-5)
- Identify critical: Google algo updates, penalty risks, major industry news
- Extract GEO/AI-search signals: AI citation readiness, brand-authority/mention trends, AI crawler analysis, platform-specific optimization (ChatGPT / Perplexity / Gemini / Google AI Overviews), schema markup for AI, and citability scoring insights
- Write analysis to
P4-cortex/growth/seo/analyzed/YYYY-MM-DD-hash.json
- If critical insight found (including GEO/AI-search shifts): create kanban task for human attention
- State file:
P4-cortex/growth/seo/seo_analysis_state.json
- LLM 필요?: ✅ Yes
3. 주간 트렌드 리포트 (Weekly Trend Report)
- Trigger cron:
seo-trend-trigger (LLM, weekly Mon 14:00 KST, fast model)
- Worker: Kanban worker
- Input: 분석된 기사 중 최근 7일치
- Process:
- Read analyzed articles from past week
- LLM synthesizes hot topics, trends, notable changes
- Write report to
P4-cortex/growth/seo/weekly/YYYY-MM-DD-trends.md
- Deliver summary to Discord channel
- LLM 필요?: ✅ Yes
4. 정리 (Cleanup)
- Script: (no_agent, monthly)
- Process: Move articles older than 90 days to
_archive/
- LLM 필요?: ❌ No
Cron Jobs Summary
| Job | Schedule | Type | Model | LLM? |
|---|
| SEO Article Harvester | 0 */6 * * * | no_agent | — | ❌ |
seo-analyze-trigger | 0 12 * * * | LLM agent | opencode-go/deepseek-v4-flash | ✅ |
seo-trend-trigger | 0 14 * * 1 | LLM agent | opencode-go/deepseek-v4-flash | ✅ |
Directory Layout
P2-hippocampus/knowledge/seo-articles/
├── YYYY/ # 연도별 수집 기사
│ └── *.md
├── _new/ # 최근 수집 (미분류)
├── _archive/ # 오래된 기사
├── collected_urls.json
└── report.json
P4-cortex/growth/seo/
├── seo_analysis_state.json # 분석 상태 추적
├── analyzed/ # 분석 결과 (YYYY-MM-DD-hash.json)
└── weekly/ # 주간 트렌드 리포트
└── YYYY-MM-DD-trends.md
RSS 피드 목록 (2026-06-05 갱신, 28개)
SEO / 마케팅 (10개) — 모든 글 통과
ahrefs.com/blog/feed, searchengineland.com/feed, semrush.com/blog/feed
yoast.com/feed/, seopress.org/feed, ascentkorea.com/feed
growthmk.com/feed, moz.com/blog/feed, searchenginejournal.com/feed
copyblogger.com/feed
테크 / 개발 (11개) — 제목 기반 키워드 필터
blog.google/technology/ai/rss, developers.googleblog.com/feeds/posts/default
blog.chromium.org/feeds/posts/default, techcrunch.com/feed
theverge.com/rss/index.xml, wired.com/feed/rss, zdnet.com/rss.xml
feeds.arstechnica.com/arstechnica/index, css-tricks.com/feed
smashingmagazine.com/feed, aws.amazon.com/blogs/aws/feed, github.blog/feed
글로벌 / 커뮤니티 (7개) — 제목 기반 키워드 필터
openschema.co.jp/feed, hnrss.org/frontpage, dev.to/feed
news.ycombinator.com/rss, techmeme.com/feed.xml
GEO/AI-Search Signal Extraction
The analyzer should surface Generative Engine Optimization (GEO) concepts alongside traditional SEO:
| Signal | What to capture |
|---|
| Citability scoring | Passage length, self-contained facts, direct-answer structure, citation readiness |
| AI crawler analysis | New AI bot signals (GPTBot, ClaudeBot, PerplexityBot), robots.txt guidance, llms.txt relevance |
| Brand authority | Brand mentions as an AI visibility signal, entity authority shifts |
| Platform optimization | ChatGPT, Perplexity, Gemini, Google AI Overviews-specific tactics |
| Schema for AI | JSON-LD, llms.txt, ARD/AI discovery signals, structured data for LLM consumption |
Analysis Format (worker output)
{
"article_url": "https://...",
"title": "Article Title",
"summary": "1-2 sentence LLM summary",
"topic_tags": ["google-update", "core-web-vitals", "ranking", "geo", "ai-search"],
"relevance_score": 4,
"geo_relevance_score": 3,
"ai_citation_readiness": 3,
"key_insight": "Main actionable takeaway",
"geo_insight": "AI-search / GEO takeaway, if any",
"is_critical": false,
"analyzed_at": "2026-06-14T12:00:00"
}
Related
- [[@action/skills/SKILL-INDEX]]
- [[@memory/growth/seo/index]]
references/crawling-tools.md — 크롤링 도구
references/discord-webhook-limits.md — Discord embed 제한