x-voice
Skill1 starsChanged 5 days agoMITView on GitHub
Measure the user's actual X voice from their own past posts and write voice.md from evidence, then check any draft for drift before it goes out. Use when the user says "write my voice.md", "learn how I write", "does this sound like me", "this doesn't sound like me", "make it sound more like me", sets up the pack for the first time, or pastes a batch of their own posts.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- liftli-ai/x-agent-skillstheir own version, first seen Oct 17, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- liftli-ai/x-agent-skills★ 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.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Every skill in this pack reads ~/.claude/x/voice.md.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
x-voice
Every skill in this pack reads ~/.claude/x/voice.md. Filled in by hand, it
says how the user would like to sound. Measured from their posts, it says how
they do sound. On X that gap is visible in one line: someone who writes in
lower-case with no full stops gets a draft in sentence case with a semicolon,
and the replies notice before the user does.
One tool lives in this folder and it runs:
python3 voice.py samples.txt # measure, print the profile
python3 voice.py samples.txt --save ~/.claude/x/voice.json
python3 voice.py --profile ~/.claude/x/voice.json --check draft.txtInput
Ten to twenty of the user's own posts, one file, --- on its own line between
posts. Posts they wrote themselves, not ones a tool wrote. Replies count too,
and on X they are often the truest sample of how someone talks.
Fewer than five and the tool says so. Do not pad the sample with posts you drafted for them: that measures you.
What it measures
Sentence length and spread, contractions, I / we / you, questions and exclamations, casing (the one that matters most on X), emoji, em dashes, numbers, length in words and lines, the words that recur across posts, and how posts open. Each is a mean and a spread, so a draft is judged against the user's own range.
Then write voice.md
voice.py samples.txt --save ~/.claude/x/voice.json- Fill
templates/voice.mdfrom what the user tells you: audience, positions, niche accounts, off limits, proof. - Paste the Measured voice block under "What I sound like on X". Where the hand-written answers disagree with the measurement, say so. The measurement wins unless the user is changing on purpose.
- Save to
~/.claude/x/voice.md.
Checking a draft
--check scores 0-100 against the profile, names the metric furthest from the
user, and raises hard flags: emoji or an em dash when they never use one,
sentence case when they write lower-case, no contractions when they always use
them. /x-preflight runs this automatically when the profile exists.
Under 25 words, sentence rhythm cannot be judged and sits out as n/a, the same
way /x-human handles short posts.
Say this honestly
Stylometry on a small sample. It catches drift. It cannot tell you a post is good, and it is not a way to imitate someone else: run it on the user's own writing only.
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.

