refine-prose
Skill4 starsChanged 6 months agoMITView on GitHub
Polishes drafted prose for clarity, conciseness, and consistent voice before it is posted or presented. Apply after drafting any plan, review summary, or written output to catch filler phrases, passive constructions, and inconsistent terminology. Use when the user asks to polish, clean up, improve, edit, or tighten writing — also triggered by "refine this draft", "make this clearer", or "improve my writing". Also composable: other skills invoke it as a late-stage refinement step. TRIGGER when: user asks to polish, edit, or improve prose. DO NOT TRIGGER when: user asks to format or lint code (use language-specific tools instead), or when this skill is already running as a composable step within another skill.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- petekp/claude-code-setuptheir own version, first seen Sep 23, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- petekp/claude-code-setup★ 48same 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 repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- audit-skillsSkill
- create-commitSkill
- format-review-commentsSkill
- implement-planSkill
- review-agents-mdSkill
- review-planSkill
- review-prSkill
- review-skillSkill
- upsert-agents-mdSkill
- upsert-planSkill
- upsert-skillSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 9 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈9
Distinctive · estimate
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 26
After drafting a plan, proposal, or structured document
- Distinctive: line 27
After writing a PR review summary or set of review comments
- Distinctive: line 28
Before presenting any intentionally drafted output — a plan, review, proposal, or multi-sentence response — to a user
- Distinctive: line 37
Remove words and phrases that add length without meaning.
- Distinctive: line 62
Remove redundant restatements: if the same idea appears twice within a paragraph, keep the clearer instance.
- Distinctive: line 63
Ensure each sentence has one main idea.
- Distinctive: line 67
All items in a bulleted or numbered list must use the same grammatical form.
- Distinctive: line 69
If the first item starts with a verb ("Run the tests"), all items must start with a verb.
- Distinctive: line 70
If the first item is a noun phrase ("Test coverage"), all items must be noun phrases.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Refine Prose
Polishes a draft for clarity, conciseness, and consistent voice. Apply after producing any written output — a plan, review summary, comment, or document — before presenting or posting it.
When to apply
- After drafting a plan, proposal, or structured document
- After writing a PR review summary or set of review comments
- Before presenting any intentionally drafted output — a plan, review, proposal, or multi-sentence response — to a user
- Whenever explicitly invoked via
/refine-proseon arbitrary text
Instructions
Work through each check in order. Apply fixes silently — do not narrate the edits. Apply prose rules only to prose — skip code blocks, inline code, tables with technical data, and structured output such as YAML frontmatter. When no changes are needed, return the original text unchanged.
1. Cut filler phrases
Remove words and phrases that add length without meaning. Common culprits:
| Remove | Replace with |
| "in order to" | "to" |
| "due to the fact that" | "because" |
| "at this point in time" | "now" |
| "it is worth noting that" | (delete or keep the fact) |
| "please note that" | (delete; just state the fact) |
| "as mentioned above/below" | (delete or use a direct reference) |
| "basically", "essentially", "literally" | (delete unless they carry meaning) |
| "very", "really", "quite", "rather" | (delete or strengthen the noun/verb) |
2. Prefer active voice
Rewrite passive constructions as active where the actor is known and the active form is clearer.
- "The tests were run by CI" → "CI ran the tests"
- "The config will be updated" → "Update the config" (in instructions)
Passive is acceptable when the actor is unknown, unimportant, or when convention demands it (e.g. "was introduced in v2").
3. Tighten sentences
- Break sentences longer than ~30 words into two unless the length is necessary for precision.
- Remove redundant restatements: if the same idea appears twice within a paragraph, keep the clearer instance.
- Ensure each sentence has one main idea.
4. Enforce parallel structure in lists
All items in a bulleted or numbered list must use the same grammatical form.
- If the first item starts with a verb ("Run the tests"), all items must start with a verb.
- If the first item is a noun phrase ("Test coverage"), all items must be noun phrases.
5. Standardize terminology
Within a single document, use one term for each concept consistently. Use whichever term already appears more often; apply the defaults below only when no term has been established yet.
Default preferences:
- "pull request" in formal prose; "PR" is acceptable in informal or space-constrained contexts
- Oxford comma in lists of three or more items
- Sentence case for headings unless the document already uses title case throughout
6. Match formality to context
- Plans and proposals: formal, third-person where possible, no contractions
- Review comments: direct and specific; second person ("you") is appropriate; contractions acceptable
- Instructions: imperative mood, second person implicit ("Run tests", not "You should run tests")
Composability
This skill is designed to be referenced by other skills as a late-stage refinement step. When wired in as a composable step, run it silently and present only the refined output — do not announce that you are running it.
Callers reference it with a line such as:
Apply the
refine-proseskill to the draft before presenting.
Examples
Direct invocation on user-supplied text:
/refine-prose
The reason that this is basically very important is due to the fact that the tests were being run by CI on a nightly basis, and it is worth noting that failures were being ignored.Refined output:
This matters because CI runs the tests nightly and ignores failures.(Applied: cut filler phrases "basically", "due to the fact that", "it is worth noting that"; rewrote passive "were being run by CI" and "were being ignored" as active.)
Composable invocation (from another skill):
Apply the `refine-prose` skill to the review summary before presenting it. Run it silently — do not announce the refinement step.When no changes are needed: Input is already clear and concise — return the original text unchanged with no commentary.
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.

