Stats
Actions
Tags
From qe-framework
Refines text for clarity and conciseness using Strunk's principles and removes AI-generated patterns. Useful for documentation, commit messages, error messages, reports, and UI text.
How this skill is triggered — by the user, by Claude, or both
Slash command
/qe-framework:Qwriting-clearlyThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Write clearly and with force. This skill covers what to do (Strunk) and what to avoid (AI patterns).
Write clearly and with force. This skill covers what to do (Strunk) and what to avoid (AI patterns).
Use this skill any time you are writing for a human:
If you are writing a sentence a human will read, use this skill.
When context is limited:
Loading a single section (~1,000–4,500 tokens) instead of full content saves significant context.
William Strunk Jr.'s The Elements of Style (1918) teaches how to write clearly and cut ruthlessly.
Elementary rules of usage (grammar/punctuation):
Elementary principles of composition:
| Section | File | ~Tokens |
|---|---|---|
| Grammar, punctuation, comma rules | 02-elementary-rules-of-usage.md | 2,500 |
| Paragraph structure, active voice, concision | 03-elementary-principles-of-composition.md | 4,500 |
| Titles, quotations, formatting | 04-a-few-matters-of-form.md | 1,000 |
| Word choice, common errors | 05-words-and-expressions-commonly-misused.md | 4,000 |
For most tasks, 03-elementary-principles-of-composition.md is sufficient. It covers active voice, positive statements, specific language, and cutting needless words.
LLMs tend to regress toward statistical averages, producing cliched and bloated prose. Avoid:
Don't write grandly — describe concretely what is actually happening.
For deeper research on why these patterns occur, see signs-of-ai-writing.md — a guide developed by Wikipedia editors to detect AI-generated submissions, well-documented and field-validated.
When writing for humans, load the relevant section from elements-of-style/ and apply the rules. For most tasks, 03-elementary-principles-of-composition.md contains the most important material.
Based on Wikipedia's "Signs of AI Writing" (WikiProject AI Cleanup). Use this section when editing text that sounds robotic, overly formal, or AI-generated.
When humanizing text:
Flat, voiceless prose is as detectable as AI-generated text. Signs of lifeless writing: uniform sentence length/structure, no opinions, no uncertainty, no first person, no humor or edge, reads like Wikipedia or a press release.
How to add voice: have opinions, vary rhythm, acknowledge complexity, use "I" where appropriate, allow messiness, be specific about feelings.
1. Inflated Significance / Legacy / Broad Trends Watch words: symbol of, testament to, pivotal/crucial role, underscores the importance, enduring legacy, evolving landscape Fix: Remove inflated framing. State what actually happened.
2. Inflated Notability / Media Coverage Watch words: independent coverage, regional/national media outlets, active social media presence Fix: Replace vague notability claims with specific citations.
3. Superficial Analysis via -ing Phrases Watch words: emphasizing, highlighting, showcasing, ensuring, reflecting, symbolizing, contributing, fostering Fix: Remove dangling -ing clauses; state the actual fact or source.
4. Promotional / Advertising Language Watch words: boasting, vibrant, rich (figurative), profound, groundbreaking, renowned, breathtaking, nestled in Fix: Replace promotional adjectives with concrete facts.
5. Vague Attribution / Hedged Claims Watch words: according to industry reports, observers noted, experts argue, some critics claim Fix: Name the specific source and date.
6. Outline-Style "Challenges and Future Outlook" Watch words: despite ... faces several challenges, future outlook Fix: Replace formulaic sections with specific facts and dates.
7. Overused AI Cliches: additionally, furthermore, delve, foster, garner, underscore, landscape (abstract), pivotal, showcase, tapestry, testament, invaluable, vibrant — replace with plain language or remove.
8. Copula Avoidance: serves as, stands as, represents, boasts, features — use simple "is/are/has."
9. Negative Parallelism: "Not only... but also..." — state the point directly.
10. Rule of Three Overuse: Don't force groups of three. Use the actual number of items.
11. Elegant Variation (Synonym Cycling): Don't swap synonyms for the same concept. Repeat the word.
12. False Range Expressions: Don't use "from X to Y" when X and Y aren't on a meaningful scale.
13. Em-Dash Overuse: Replace most em-dashes with commas, periods, or parentheses.
14. Bold Overuse: Remove mechanical bolding of terms that don't need emphasis.
15. Inline Header Lists: Convert bold-header bullet lists into flowing prose.
16. Title Case Overuse: Use sentence case for headings.
17. Emoji: Remove decorative emoji from headings and list items.
18. Curly Quotes: Replace curly quotes with straight quotes.
19. Collaborative Communication Artifacts: Remove chatbot response phrases (I hope this helps, Of course!, Certainly!, Would you like, Let me know, Here is).
20. Training Data Cutoff Disclosures: Replace "as of my knowledge cutoff" with actual sources and dates.
21. Sycophantic Tone: Remove excessive praise. State the substance.
22. Filler Phrases: "In order to achieve this goal" → "To achieve this" / "Due to the fact that" → "Because" / "has the ability to" → "can" / "It is important to note that" → (delete)
23. Excessive Hedging: Reduce qualifiers. "could potentially might have some influence" → "may influence outcomes"
24. Vague Positive Endings: Replace bland optimism with specific plans.
Wikipedia: Signs of AI Writing — Core insight: "LLMs use statistical algorithms to predict what comes next. The result tends to converge on the most statistically plausible outcome for the broadest set of cases."
npx claudepluginhub inho-team/qe-framework --plugin qe-frameworkCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.