blog-post

name	blog-post
description	Create a blog post

Blog Post Writer

Create a well-researched blog post on a given topic. Researches the area, writes without em dashes, and adds a validated References section with external links.

Process

Follow these steps in order:

1. Research the Topic

Use the web-search skill to research the topic thoroughly:

uv run {baseDir}/../web-search/search.py "<topic>" -n 10

Read the top results to gather facts, perspectives, and sources:

uv run {baseDir}/../web-search/content.py "<url>"

Gather at least 3 high quality external sources to reference.

2. Write the Blog Post

Write the blog post as a markdown file. Follow these rules strictly:

• Never use em dashes (—) or en dashes (–). Use commas, periods, or parentheses instead. Regular hyphens are fine in compound words like "counter-drone", "anti-aircraft", "long-term".
• Write like a human, not an AI. Specifically:
  • Use simple, everyday words. Write "use" not "utilize", "help" not "facilitate", "start" not "embark", "show" not "showcase", "change" not "revolutionize".
  • Ban these AI slop words/phrases entirely: "delve", "tapestry", "landscape", "paradigm", "leverage", "robust", "comprehensive", "cutting-edge", "groundbreaking", "game-changing", "pivotal", "seamlessly", "furthermore", "moreover", "it's worth noting", "in today's world", "at its core", "in the realm of", "the power of".
  • Keep sentences short. If a sentence has more than 25 words, split it.
  • Write in active voice. Not "the results were analyzed" but "we analyzed the results".
  • Be direct. State the point first, then support it. No flowery introductions.
  • Vary sentence length and structure. Monotone rhythm = AI smell.
  • Write at a 8th grade reading level. If a simpler word exists, use it.
• Use proper markdown headings, paragraphs, and formatting
• No generic "compelling" introductions or "in conclusion" wrapups. Start with the most interesting fact or claim. End when you're done.

3. Add Wikipedia Links for Technical Terms

Identify 5 to 10 key technical terms in the blog post (acronyms, named technologies, concepts, algorithms, etc.). Resolve their Wikipedia URLs:

uv run {baseDir}/wikipedia_links.py '["term1", "term2", "term3"]'

The script returns a JSON object mapping each term to its Wikipedia URL (or null if not found). For each term with a resolved URL, add an inline link on its first occurrence in the post body. Do not link the same term twice. Skip terms where the URL is null.

Example: replace REST API with [REST API](https://en.wikipedia.org/wiki/REST).

4. Add a References Section

At the end of the post, add a ## References section with numbered external links. Each reference must:

• Link to a real, accessible URL
• Have anchor text that accurately describes the linked content
• Be relevant to the blog post topic

Example:

## References

1. [Understanding Modern Web Frameworks](https://example.com/web-frameworks)
2. [Performance Benchmarks 2025](https://example.com/benchmarks)

5. Validate All Reference Links

After writing, validate every link in the References section:

uv run {baseDir}/validate_links.py '[{"url": "https://...", "anchor": "link text"}, ...]'

Pass all reference links as a JSON array. Each entry needs a url and anchor field.

If any link fails validation:

• Not reachable: Find an alternative URL for the same topic
• Content mismatch: Update the anchor text to match the actual page content, or find a better URL

Re-validate until all links pass.

Output

Save the blog post as a markdown file in the current working directory. Use a slug derived from the title as the filename (e.g., understanding-web-frameworks.md). Note: hyphens are allowed in filenames, just not in the blog post content itself.

Dependencies

This skill requires the web-search skill to be available at {baseDir}/../web-search/.