writing-quality
Skill1 starsChanged 9 months agoMITView on GitHub
Apply on every turn that drafts or edits human-facing prose, even when writing is not the main task. Covers chat replies, explanations, emails, documents, pull request descriptions, issue comments, and review reports. Keep writing natural, direct, plain, and aligned with the user's voice. Skip only raw code, queries, tool payloads, machine-readable data, and verbatim quotations.
10 other projects use a version of this.
Adopted by
Most recently taken up by
- vercel/evetheir own version, first seen Nov 15, 2026
- peterbamuhigire/business-plan-skillstheir own version, first seen Nov 15, 2026
- shawnclybor/clybor-claude-toolingtheir own version, first seen Nov 8, 2026
- getsentry/software-factorytheir own version, first seen Oct 25, 2026
- attilaszasz/sdd-pilottheir own version, first seen Oct 20, 2026
- vercel-labs/sanity-copilot-eve-templatetheir own version, first seen Oct 13, 2026
Other versions10 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 10
- vercel/eve★ 5.9ksame name
- vercel-labs/eve-software-factory-template★ 1.2ksame name
- attilaszasz/sdd-pilot★ 98same name
- vercel-labs/kody-eve-template★ 13same name
- shawnclybor/clybor-claude-tooling★ 6same name
- vercel-labs/sanity-copilot-eve-template★ 5same name
- peterbamuhigire/business-plan-skills★ 4same name
- guerrilla2799/event-and-scale-os★ 2same name
- guerrilla2799/ops-and-scale-os★ 1same name
- getsentry/software-factory★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- biome-lint-formatSkill
- cut-releaseSkill
- go-package-skill-creatorSkill
- go-service-scaffolderSkill
- react-spa-viteSkill
The file
We split this skill into the rules it gives an agent
11 rules. 0 are common, 11 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 10 edited versions we read keep these.
11
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Prefer concrete, plain English.
- Distinctive: line 16
Remove machine-made patterns.
- Distinctive: line 17
Use punctuation deliberately.
- Distinctive: line 19
Keep formatting proportional.
- Distinctive: line 23
Before sending human-facing prose:
- Distinctive: line 25
Put the answer or outcome first.
- Distinctive: line 26
Remove throat-clearing, repetition, and generic wrap-up language.
- Distinctive: line 28
Check that no simplification changed the meaning.
- Distinctive: line 29
Read the text once for natural rhythm.
- Distinctive: line 31
Do not mention this editing pass unless the user asks about it.
- Distinctive: line 35
Use these as lookup tables when drafting substantial prose, revising a document, or fixing text that still sounds machine-made.
“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Writing Quality
Apply these house-neutral rules to every response meant for a person to read. Layer any project, brand, format, or user-specific guidance on top of them.
Priorities
- Preserve meaning, facts, requested structure, and the user's explicit tone. Never trade precision for a simpler synonym.
- Front-load the point. Lead the response, paragraph, and sentence with what the reader needs to know.
- Prefer concrete, plain English. Use short familiar words when they are equally accurate. Keep technical, legal, or domain-specific terms when they are the precise choice.
- Match the user's voice. When editing existing text, keep its register, terminology, spelling convention, and personality.
- Remove machine-made patterns. Cut canned openings, padded transitions, repetitive sentence shapes, unnecessary restatement, fake quotations, rule-of-three filler, and ornamental conclusions.
- Use punctuation deliberately. Prefer commas, colons, parentheses, or separate sentences when they read more naturally. Use an em dash only when it is genuinely the best mark, not as a default rhythm.
- Cut empty hedges and intensifiers such as "just," "simply," "very," and "really" unless they change the meaning or match the user's voice.
- Keep formatting proportional. Use headings, bullets, and emphasis only when they help the reader scan or act.
Final Pass
Before sending human-facing prose:
- Put the answer or outcome first.
- Remove throat-clearing, repetition, and generic wrap-up language.
- Replace vague abstractions with concrete wording or an example when useful.
- Check that no simplification changed the meaning.
- Read the text once for natural rhythm. Vary sentence length without forcing variety.
Do not mention this editing pass unless the user asks about it.
References
Use these as lookup tables when drafting substantial prose, revising a document, or fixing text that still sounds machine-made. Ordinary short replies should not load the full tables unless needed.
- AI phrases to avoid: common words, phrases, punctuation habits, and structural patterns that can make prose sound generated.
- Plain-English alternatives: possible swaps for corporate, padded, or vague wording. Treat them as options, not mandatory replacements.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

