| name | content-production |
| version | 1.0.0 |
| description | Full content production pipeline — takes a topic from blank page to published-ready piece. Use when you need to execute content: write a blog post, article, or guide end-to-end. Triggers: 'write a post about', 'draft an article', 'create content for', 'help me write', 'I need a blog post'. NOT for content strategy or calendar planning (use content-strategy). NOT for repurposing existing content (use content-repurposing). NOT for social captions only. |
| triggers | ["content production","content pipeline","content workflow"] |
When to Use
Use when executing a piece of content end-to-end: a blog post, article, or guide from blank page to publish-ready draft. If the question is what to publish or when, use content-strategy. If the task is de-AI-ifying existing copy, use content-humanizer. If the request is social captions only, use social-content.
Workflow
-
Brief — 6 questions before writing anything. If none exists, ask or infer and state assumptions: the reader and their single job; the one thing they should believe or do after reading; format (how-to, listicle, opinion, comparison, case study); where it lives (blog, newsletter, guest post); proof available (data, quotes, examples); target keyword or search intent, if SEO-relevant.
-
Research. Gather 3–5 credible sources. Pull specifics: numbers, named examples, tool names, named experts. Unverifiable? Flag it — never invent stats, quotes, or case results.
-
Outline. Write an H2-level outline with a working title, a one-sentence thesis, and the key claim under each section. Each H2 answers a question the reader actually asked. For pieces over ~1,500 words or high-stakes work, get approval before drafting.
-
Draft. Follow the outline; write the opening last if helpful. Paragraphs under 4 lines, one idea each, answer up front (inverted pyramid).
-
Edit passes — three separate passes, in order:
- Structure: does every section serve the thesis? Cut what doesn't. One H1, logical H2s, H3s only under H2s. Move the strongest point earlier.
- Clarity: every sentence clear on first read? Kill jargon, passive voice, hedges ("somewhat", "arguably"). Replace abstractions with concrete examples.
- Voice: read it aloud — does it sound like a person? Vary sentence length, use contractions, add first-person judgment, strip AI tells (see content-humanizer).
-
SEO pass (only if targeting search). From the finished draft:
- Title tag: primary keyword near the front, under 60 chars, a reason to click.
- H1: matches reader intent, not necessarily the title tag.
- Meta description: under 155 chars, keyword + payoff.
- Keywords: one primary, 2–3 secondary, placed naturally — no stuffing.
-
Publish-ready checklist. All must pass before declaring done:
Draft-Review Rubric
Score 1–5 (5 = ready to publish). Fix anything below 4 before shipping.
| Criterion | What 5 looks like |
|---|
| Thesis | Reader can state it after a skim |
| Structure | Every section advances the thesis |
| Clarity | Every sentence clear on first read |
| Specificity | Concrete numbers/names/examples throughout |
| Voice | A knowledgeable person, not a committee |
| Accuracy | Claims verified; unknowns flagged |
| SEO (if applicable) | Title/H1/meta/keywords set and natural |
| CTA | One clear next step tied to reader intent |
Pitfalls
- Writing before the brief: a wrong-target piece costs more than the brief ever does.
- Outlines that mirror topic taxonomy instead of reader questions — restructure around search intent.
- Faking proof: invented stats, quotes, or case studies kill credibility; mark unverifiable claims.
- Editing for voice before structure — polishing prose you'll cut wastes the pass.
- Writing meta/title first, then forcing the draft to fit; SEO belongs at the end.
Verification
- Confirm the step 7 checklist is fully ticked.
- Re-read cold: can a stranger state thesis, audience, and CTA after one skim?
- If a brief existed, diff the piece against it — every brief question answered?
- Run the draft-review rubric and report scores alongside the piece.