SEO Article Harvester

SEO 트렌드 + AI 기사를 자동으로 수집/스코어링/큐레이션하는 cron pipeline. humanerd.kr에 적합한 evergreen 콘텐츠 소스를 발굴한다.

작동 흐름

RSS feed (~20 sources)                  [매 6시간 cron]
  ↓ feedparser
collected/YYYY-MM-DD/*.json              ← raw collected
  ↓ scoring (0~10)
scored/YYYY-MM-DD/*.json                ← with score
  ↓ threshold ≥ 7
analyzed/keep/*.json                     ← high-quality keeper
  ↓
  └→ human review (Obsidian)
  └→ content-pipeline (kanban task)
  └→ weekly digest (Trend Harvester)

Cron

• SEO Article Harvester: 0 */6 * * * (6h cycle)
• Trend Harvester: 0 */6 * * * (6h cycle)
• content-pipeline: 0 */3 * * * (3h cycle, 현재 disabled)

Source categories

• AI/ML research: OpenAI blog, Anthropic news, Google DeepMind, Hugging Face
• Korean tech: 카카오 AI, 네이버 D2, 당근마켓 블로그
• AI dev tools: GitHub trending (AI), Hacker News (filtered)
• Newsletter: TLDR AI, Ben’s Bites, Import AI

Scoring (0~10)

• Relevance to humanerd.kr (agent/automation/devtools) — 40%
• Recency (last 30 days) — 20%
• Source authority (Tier 1/2/3) — 20%
• Content depth (>1000 words, primary source) — 20%

통계 (2026-06-01)

• 175+ articles collected (Trend Harvester)
• 35+ articles kept (SEO Harvester, last 30 days)
• 6 articles promoted to “human review needed”
• 1 article in “ready to publish” (status: publish)

관련

• skills/seo-article-harvester — cron job skill
• services/seo-harvester — product detail page
• drewgent — Drewgent 본체