Give every article a claim ledger before you give it a hook

Produce a useful sourced article through distinct drafting, factual editing and style-review passes.

SeriesField notes
SubjectClaude skill
Issued
Revised

Field notes. John Arndt, Soxoa. Published .

The sentence that makes an article more interesting is often the sentence its sources cannot support. A small product test becomes a market trend. Search impressions become demand. A proposed customer workflow becomes a case study.

The fix starts before the first paragraph. Write down what the evidence actually establishes, then build the article around that.

In September 2026, we used this approach for an article about local AI visibility. The retained work includes source notes, successive drafts, editorial reviews and a final article. The writing moved through copywriting, factual editing and a separate style review. The final record described a completed draft with no unresolved material editorial issues. It did not claim the article had been published.

That workflow is reusable because the difficult decisions repeat across subjects.

Keep the claim attached to its measurement

A claim ledger can be a small table. Record the claim, the source, the date, what was measured and the limitation that needs to travel with it.

Suppose an illustrative business test contains twenty sampled assistant answers. Three name the business. You can describe the result in that sample. You cannot turn it into customer reach, market share or a universal rank. The questions, models, location and date affect what the result means.

The same discipline applies elsewhere. A Search Console impression is an appearance in a particular reporting context. It is not automatically monthly search volume. A product demonstration shows behavior under its test conditions. It does not establish that a customer saved ten hours.

These distinctions should survive the editing process. A stronger hook is not a reason to delete the denominator.

Give drafting and review different jobs

The writer's job is to make the useful point understandable. The editor's job is to challenge the point, the structure and the support. A style reviewer looks for repeated instructions, generic openings, vague abstractions and a voice that sounds assembled from familiar formulas.

Those can be separate people or separate passes by one assistant. Independence can help when the stakes justify it. The useful part is giving each pass a clear responsibility and retaining the issues it found.

In the September article, the source notes explicitly rejected attractive claims that had arrived in supplied reference material without independent verification. The article could still borrow a useful structure: answer a commercial question, explain the method, show original evidence and give the reader a practical next step. It did not need to inherit the reference's unsupported statistics.

That is a productive editing rule. Keep the useful idea while checking its factual cargo.

Stop revising when the material issues are resolved

A repeatable workflow needs a stopping condition. Otherwise an assistant can keep replacing clear sentences with different clear sentences indefinitely.

Resolve the factual and structural issues, check the changed passages and confirm that the reader can act on the explanation. Record remaining limitations. A missing fact can require a narrower conclusion rather than another round of adjectives.

Style review also needs an honest label. It can identify formulaic writing. It cannot prove who wrote a passage, and an invented AI-detection score adds no evidence.

The new Claude skill draft, write-evidence-article, turns this process into portable instructions. Give it the reader's question, intended publisher, source material and voice requirements. It returns a finished article with source and editorial notes where useful. The historical process produced a reviewed article; the newly packaged Claude version has not yet been evaluated in Claude.

A useful first request is:

“Use write-evidence-article to explain this workflow to a small-business owner. Build a claim ledger from the attached sources. Keep observations, estimates and illustrative examples distinct. Review the draft for factual scope and formulaic writing before returning it.”

Choose a subject where you already have a few solid sources. Read the final draft beside the ledger. If the article makes a stronger claim than the evidence, change the article. That is the review that matters.

Get the skill

write-evidence-article is part of the free Soxoa Ops plugin for Claude. In Claude Code, add the marketplace with /plugin marketplace add johnarndt/soxoa-claude-plugins, then run /plugin install soxoa-ops@soxoa-plugins. You can read the skill on GitHub before installing it, or copy its folder into ~/.claude/skills/.

Bring the workflow you think is worst.

Thirty minutes, with me, and an honest answer about whether there is anything worth building. Scope and price are agreed before work begins.