All skills
simota avatar

/zine

@8e1f365
by shingo imotasimota/agent-skills85 stars
15

Authoring tech blog/article series for note/Zenn/Qiita/dev.to. Not for specs (Scribe) or microcopy (Prose).

Use this Skill: https://skilld.dev/gh/simota/agent-skills/zine

This session only. Nothing lands on disk.

referencecontent-repurposing.md

≈2.7k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Content Repurposing

Purpose

One canonical article should become many deliverables across platforms — note, Zenn, Qiita, dev.to, X threads, LinkedIn posts, newsletter, and atomic social cards — without lossy translation or duplicate-content penalties. Repurposing multiplies reach 5–10× per hour invested compared to writing fresh.

Scope Boundary

  • IN scope: canonical-to-variant transformation, atomic asset extraction (threads, quote cards, snippets), platform-specific re-tone, canonical URL strategy, scheduling cadence.
  • OUT of scope: original article authoring (note/zenn/qiita/devto), SEO duplicate-content audit (delegate to growth), social ad creative (different domain), video cuts (delegate to cue/director), thumbnail design (delegate to director).

Core Concepts

Hub-and-Spoke Model

Treat one platform as the hub (canonical home) and the others as spokes (variants linking back).

Hub Choice When
Personal blog / company blog When SEO authority should accrue to your own domain
note JP-first, magazine series, paid content
Zenn Engineer-first, code-heavy, OSS context
dev.to English global reach, cross-post via canonical_url

Spokes use canonical_url (dev.to / Medium support this natively) or a "originally posted at" link to avoid duplicate-content SEO penalty.

The 1 → 5 → 25 Multiplier

A single 3,000-word article should yield:

Level Count Output
1 1 Canonical long-form
5 5 Platform variants — note / Zenn / Qiita / dev.to / LinkedIn long-form
25 20+ Atomic assets — X thread, LinkedIn post, quote cards (5–10), code snippets, single-tip posts, newsletter blurb, Bluesky post, Mastodon share, podcast talking point, video script outline

Medium note (2025-2026): Medium Partner Program has undergone multiple paywall/earnings restructurings since November 2025 — search-driven traffic, new-member conversion, and broader distribution to non-Boosted stories now all affect earnings. Ongoing eligibility requires 5+ stories published within the last 6 months. Source: medium.com Partner Program blog. Evaluate Medium as a spoke carefully; canonical_url is supported via "Import a story". For JP-first content, Medium is generally lower priority than Zenn/Qiita.

Don't aim for all 25 every time; aim for the highest-leverage 8–10 per article.

Variant Translation, Not Copy-Paste

Real repurposing adapts:

Dimension Adaptation
Length note 4000字 → dev.to 1500 words is not just translation; it's a re-pacing
Voice note 一人称で良い → Qiita では具体手順中心
Code samples dev.to may include English comments; Zenn keeps Japanese; note often hides verbose code in 折りたたみ
Hook Hook reset for each platform's audience expectation
CTA note → "次回記事を購読" / dev.to → "follow me on X" / LinkedIn → "what's your experience?"
Title Per-platform tone (see headline-patterns.md)
Length budget Each platform's no-ellipsis title cap
Tags Different taxonomy per platform
Image OGP / cover image / thumbnail varies per platform

Atomic Asset Catalog

Asset From Format
X thread First 6–10 sentences of canonical Tweet 1 = hook, 2–N = beats, last = CTA + canonical link
LinkedIn post Listicle section or contrarian beat 3 lines hook, body, "what's your take?" close
Quote card One-sentence insight 1080×1080 image, branded, no logo dominance
Bluesky / Mastodon Hook + canonical link < 300 chars; Bluesky MAU ~1.5-3M (as of Oct 2025, down 40% YoY); strongest among developers, journalists, researchers
Newsletter blurb TL;DR + link 60–120 words
Code snippet card Code block + 1-line caption Carbon / ray.so / shiki render
Single-tip post One paragraph from the listicle Stand-alone "did you know" framing
Podcast talking point Ledger of section headings 3–5 talking points for solo or guest spot
Video outline H2 hierarchy + voiceover lines Hand off to cue for production
Slide-flash 6–10 slides Hand off to stage

Canonical URL Strategy

Platform Native canonical support
dev.to canonical_url: in front-matter — first-class
Medium "Republished" via "Import a story" or canonical_url query param
Hashnode canonical: in front-matter
Substack Link in footer ("Originally published at …")
note No native canonical; rely on first-publish primacy
Zenn No native canonical; the URL hierarchy itself is the signal
Qiita No native canonical; uses self-hosted only

For Google ranking, the rule is: publish on the canonical hub first, wait 24–72 hours for indexing, then publish spokes with explicit canonical URL pointing back. Reverse order can flip the canonical to the wrong domain.

Cadence

Don't publish all variants the same day — saturates feeds and looks spammy.

Time Action
Day 0 Hub publishes (canonical)
Day 0 X thread teaser (4–6 hours after)
Day 1 LinkedIn post variant
Day 2 First spoke (Zenn or dev.to) with canonical pointing to hub
Day 4 Second spoke (Qiita or Medium)
Day 7 Newsletter feature
Day 14 Quote card series begins (drip across week)
Day 30 "Best of" round-up post on hub if multi-article series

Re-Engagement Without Re-Posting

For evergreen content, plan re-shares at 30 / 90 / 180 days with refreshed framing:

Refresh Type Example
Anniversary "1 year later: how this advice held up"
Counter "Why I changed my mind about X"
Application "Three teams that used this technique"
Update "2026 edition: what's new"

Each refresh is a new article, not a repost — original article gets internal link from the refresh.

Translation vs Localization

Cross-language repurposing (Japanese → English or vice versa) is localization, not translation:

  • Examples and metaphors must change (yakitori → BBQ; 麻雀 → poker).
  • Frameworks named differently (Qiita → Stack Overflow / Medium reference).
  • Tone calibration: Japanese tech audience often prefers reserved, English audience often prefers bolder claims.
  • LLM auto-translation produces detectable seams; always have a human review or rewrite key sections.

Workflow

  1. Identify the hub — which platform is the canonical home?
  2. Publish on the hub first and wait for index.
  3. Plan the variant set — pick 3–5 spokes from the platform table.
  4. Plan the atomic set — 5–10 assets from the catalog.
  5. Adapt each variant — title, length, voice, examples, CTA. Not copy-paste.
  6. Set canonical URLs explicitly on platforms that support it.
  7. Schedule with cadence — drip across 1–2 weeks.
  8. Track performance per spoke — find which platform pulls hardest for this content shape.
  9. Plan refresh at 30 / 90 / 180 days for evergreen pieces.

Output Template

repurposing_plan:
  hub:
    platform: own_blog
    canonical_url: https://example.com/posts/flaky-tests
    publish_date: 2026-04-25
  spokes:
    - platform: dev.to
      title: "Diagnose flaky tests with bisect-style isolation"
      length_target_words: 1500
      canonical_url_in_frontmatter: yes
      voice: "friendly english, slightly hooky"
      cta: "follow me on Bluesky"
      publish_date: 2026-04-27
    - platform: zenn
      title: "Flaky test を bisect で特定する実践手順"
      length_target_chars: 4500
      voice: "engineer peer, code-forward"
      cta: "コメントで実例お待ちしてます"
      publish_date: 2026-04-29
  atomic_assets:
    - type: x_thread
      tweet_count: 8
      hook_first_tweet: "テストを増やすほど信頼が下がる理由"
      publish_date: 2026-04-25T18:00
    - type: linkedin_post
      angle: "contrarian outcome lesson"
      length_chars: 1200
      publish_date: 2026-04-26
    - type: quote_card
      count: 5
      key_insights: ["...", "...", "...", "...", "..."]
      drip_schedule: "every 2 days"
    - type: newsletter_blurb
      length_words: 90
      publish_date: 2026-05-02
  refresh_plan:
    - days_post_publish: 30
      type: counter
      title: "What I'd change one month later"
    - days_post_publish: 180
      type: anniversary
      title: "Six months of bisect-driven test debugging"

Anti-Patterns

  • Copy-pasting identical text to every platform — feeds detect duplicates and downrank.
  • Publishing spokes before the hub — flips canonical to the wrong domain.
  • Skipping canonical_url on dev.to / Medium where it's native.
  • Same-day publish across all platforms — looks spammy, dilutes own attention budget.
  • Identical hook on X thread and LinkedIn — different audiences, different angles.
  • Translating idioms literally instead of localizing — alienates the target audience.
  • Quote card series with 5 cards in 1 day — same audience saturation problem.
  • Ignoring a high-CTR platform because it "feels off-brand" — let data update the brand.
  • Refreshing by re-posting the same URL — no new value; missed opportunity for an internal-link gain.

Deliverable Contract

A repurposing plan is complete when:

  • Hub platform identified with canonical URL.
  • Spoke list (3–5) with adapted title / length / voice / CTA per platform.
  • Atomic asset list (5–10) with formats and drip schedule.
  • Canonical URL strategy confirmed per platform.
  • Cadence calendar drafted (1–2 weeks).
  • Refresh plan scheduled for evergreen pieces.
  • Localization notes if cross-language.

References

  • Gary Vaynerchuk, Crushing It! (2018) — content multiplication / pillar-and-cluster framework.
  • Andy Crestodina, Orbit Media Studios — content atomization research.
  • Ross Hudgens, Siege Media — repurposing playbooks for SEO-resilient cross-posting.
  • dev.to Editor Guide — cross-posting and canonical_url field (official docs).
  • Ahrefs Blog, Duplicate Content (2024) — when canonical works and when it doesn't.
  • Sprout Social, Bluesky Growth Statistics 2026 — platform reach data for distribution decisions.
  • Semrush AI Overviews Study 2025 — canonical placement vs AI citation rates.
  • Medium Partner Program January 2026 Update — earnings model changes affecting cross-post strategy.

Source: SKILL.md on GitHub

No alerts5mo3 checks · Risk SAFE
  • Gen Agent Trust Hub5mo

    The 'zine' skill is a comprehensive instruction set for technical writing and content distribution. It contains no executable code or malicious patterns and includes proactive security guidelines for protecting sensitive data during the authoring process.

  • Socket5mo

    No alerts

  • Snyk5mo

    Risk: LOW · No issues

Signed by skilld at 8e1f365. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated last month

README badge

README badge for simota/agent-skills/zine