Writing Quality Agent Skill

Writing Quality

House-neutral rules for judging whether drafted content reads like a person wrote it. They apply to any prose surface. Voice or audience context from the caller layers on top of them.

When to Apply

Load this skill at the start of every review, before reading the draft.

Core Rules

• Kill the AI tells: em-dash overuse, "delve", "leverage", "it's not just X, it's Y", rule-of-three padding, and the rest of the patterns in references/ai-phrases-to-avoid.md.
• Prefer plain English. Swap bloated or vague wording for the shorter, concrete alternative. references/plain-english-alternatives.md is the lookup table.
• Front-load the point. Lead sentences, paragraphs, and sections with the conclusion, because readers scan.
• Concrete over abstract. Show an example before stating a principle, and cut hedges like "just", "simply", "very", and "really".
• Match the user's voice, not a default. When judging edits to existing content, respect its register and conventions. These rules trim the noise; they don't impose a personality.

References

Load the reference that matches the draft rather than all of them at once:

• references/ai-phrases-to-avoid.md: words, phrases, and punctuation patterns that mark text as AI-generated, with replacements. Applies to every review.
• references/plain-english-alternatives.md: plain-English swaps for corporate, padded, or vague wording. Applies to every review.
• references/web-writing-best-practices.md: a sourced checklist for long-form web content, covering scanning behavior, hooks, structure, readability, length by intent, and closes. Apply when the draft is an article, blog post, or guide.
• references/web-content-specs.md: the concrete numbers behind that checklist (slug rules, heading cadence, readability thresholds, word count by intent, alt text, internal links). Apply for a final spec check on long-form pieces.