blog-spark
Write an SEO-optimized teaser blog for an ebook-weaver topic — opens curiosity loops, teases 2-3 cited facts, drives readers to the ebook, and never resolves the questions it raises. Trigger the user types /blog-spark "<topic>" or /blog-spark <topic-folder-path>.
Works with
--- name: blog-spark description: Write an SEO-optimized teaser blog for an ebook-weaver topic — opens curiosity loops, teases 2-3 cited facts, drives readers to the ebook, and never resolves the questions it raises. Trigger the user types /blog-spark "<topic>" or /blog-spark <topic-folder-path>. license: MIT --- # /blog-spark Write one SEO-optimized teaser blog whose **only job** is to create curiosity so readers go read the ebook. It is not a summary and it never resolves the questions it raises — the book does that. ## Usage ``` /blog-spark "<topic>" # standalone: light research, then write /blog-spark <path-to-topic-folder> # from ebook-weaver: reuse _research/, link BOOK.md ``` ## Vault location Resolve `<vault>` — **never hardcode a personal path**: use `--vault <path>` if given, otherwise the **current working directory**. ## What You Must Do When Invoked ### Step 0 — Resolve mode - If the argument is an existing folder containing `BOOK.md` → **book mode** (`TOPIC_DIR` = that folder). - Else treat the argument as a topic string; check `<vault>/<slug>/BOOK.md`. If found → book mode. Otherwise → **standalone mode**: create `<vault>/<slug>/` with a minimal `_research/` (sources.md + fact-check.md) and do light research first (Step 1). ### Step 1 — Gather (only what's missing) - **Book mode:** read `_research/sources.md`, `_research/fact-check.md`, and skim the chapters. Do NOT re-research. - **Standalone mode:** WebSearch/WebFetch 3–5 sources, populate `sources.md` + `fact-check.md` (same ledger format as ebook-weaver). Only `verified`/`single-source` facts may be used. ### Step 2 — Write `BLOG.md` into `TOPIC_DIR` Frontmatter: ```yaml --- title: <≤60 chars, primary keyword included, curiosity-framed> description: <meta description ≤155 chars, ends on tension/question> slug: <kebab-case> keywords: [primary, secondary, secondary] date: <YYYY-MM-DD> book: "[[BOOK]]" --- ``` Body (~1,200–1,800 words): 1. **Hook intro** (100–150 w) — open with the most surprising verified fact or a scene. Primary keyword in the first 100 words. 2. **3–4 H2 sections**, each an **open loop**: pose a question the ebook answers, tease the setup, withhold the payoff. One H2 contains the primary keyword. 3. **2–3 teased facts** with citations. Footnotes must carry the **real clickable link or full reference** (URL as a markdown link, or Author/Title/Year for offline/media), not just an internal `Sx` pointer — book mode pulls these from `_research/sources.md`. 4. **Question cluster** near the end — 3–5 direct questions to the reader (mirror the ebook's closing questions, unanswered). 5. **CTA** — point to the ebook by its catchy title with an internal wikilink to `[[BOOK]]`. ### Step 3 — Verify SEO checklist (before finishing) - [ ] Title ≤60 chars, keyword present, curiosity-framed - [ ] Meta description ≤155 chars, ends on question/tension - [ ] Primary keyword in H1, first 100 words, ≥1 H2, ~4–6× total (no stuffing) - [ ] H1 → H2 → H3 hierarchy, no skipped levels - [ ] 1 image (reuse strongest from `_assets/`) with keyword-bearing alt text - [ ] Internal link to `[[BOOK]]`; external links only to cited sources - [ ] Short paragraphs (≤3 sentences), scannable - [ ] Reading level easy (~grade 7–8) ### Step 4 — Report Tell the user: title, word count, keywords, checklist pass/fail, path to `BLOG.md`. ## Anti-goals - Do NOT summarize the book or resolve its questions — sell the itch, don't scratch it. - Book mode: do NOT introduce facts absent from `fact-check.md`. - No clickbait the ebook can't cash. ## Standing rules - Never hard-delete vault notes. Do not commit/push.
More Data Engineering skills
data-pipeline
claude-office-skills/skills
Data pipeline and ETL automation - extract, transform, load workflows for data integration and analytics
ETL Pipeline
claude-office-skills/skills
Design and automate Extract, Transform, Load data pipelines for data integration and analytics
data-throughput-accelerator
affaan-m/ecc
Use when large data ingestion, backfill, export, ETL, warehouse loading, manifest catch-up, or table synchronization needs to become much faster while preserving data correctness.

