voice-interview
Skill7 starsChanged 19 days agoMITView on GitHub
Set up the writing-voice profile. Say "set up my voice profile", "set up my voice", "voice interview", or "capture my voice" (or run /founder-os:voice-interview). This is how your WRITING sounds; if they mean talking to the OS out loud (speaking and listening), that is add-voice, not this. Interactive interview that captures rhythm, openings, closings, contractions, idiosyncrasies, and writes the result to `core/voice-profile.yml`. Asks for samples first because samples beat self-description, then asks shaping questions to fill the gaps. Extracts intent from messy real-world answers, never asks the user to be structured.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- guerrilla2799/sdr-copilottheir own version, first seen Oct 20, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- guerrilla2799/sdr-copilot★ 1same 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 repository100 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- notion-package/CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- add-handsSkill
- add-mouthSkill
- add-voiceSkill
- approval-gatesSkill
- auditSkill
- backupSkill
- blind-spot-reviewSkill
- bottleneck-diagnosticSkill
- brain-logSkill
- brain-passSkill
- brain-snapshotSkill
- brand-interviewSkill
- brand-voice-interviewSkill
- business-context-loaderSkill
- campaign-from-themeSkill
- capture-sweepSkill
- catch-upSkill
- client-updateSkill
- connectSkill
- content-repurposerSkill
- content-startSkill
- context-persistenceSkill
- cross-linkSkill
- data-securitySkill
- decision-frameworkSkill
- decisions-startSkill
- delivery-startSkill
- email-drafterSkill
- employee-reviewSkill
- finance-importSkill
- forcing-questionsSkill
- founder-coachingSkill
- founder-next-moveSkill
- founder-os-setupSkill
- founder-reviewSkill
- founder-scope-challengeSkill
- github-opsSkill
- handoff-protocolSkill
- hireSkill
- historySkill
- housekeepingSkill
- ingestSkill
- knowledge-captureSkill
- legal-complianceSkill
- linkedin-brand-directionSkill
- linkedin-network-scanSkill
- linkedin-postSkill
- linkedin-power-auditSkill
- linkedin-startSkill
- linkedin-warm-revivalSkill
- lintSkill
- list-prunerSkill
- log-archiveSkill
- log-replySkill
- meeting-prepSkill
- memory-passSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
12 rules. About 0 are common, 5 are distinctive, and 7 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 24
Do not invent answers. If the user skips a question, leave the field as [NOT SET] and tell them they can re-run the interview later.
- Distinctive: line 29
Run the phases in order. When you reach a heavy phase, read its reference file for the full procedure, then execute it.
- Very specific: line 56
If they say "brand", stop and dispatch to brand-voice-interview instead.
- Distinctive: line 58
If they say "personal" or "both", continue with this skill for personal voice.
- Very specific: line 70
Read references/samples-and-questions.md.
- Very specific: line 82
Read references/confirm-and-save.md.
- Distinctive: line 94
Stop. Do not do anything else.
- Very specific: line 100
If the user runs this skill and core/voice-profile.yml already has real values (not template placeholders), ask:
- Very specific: line 104
If update, ask which fields, then walk only those questions.
- Very specific: line 105
If start-over, run the full interview.
- Distinctive: line 113
Never tell the user to be more concise.
- Distinctive: line 114
Never invent samples. Never invent vocabulary.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
Voice Interview
Runs on: local-writes - creates or edits files in your OS folder; needs an agent with write access.
You are running an interactive interview to capture the user's writing voice. The output is core/voice-profile.yml. The voice-profile feeds the your-voice skill, which then writes everything as the user from that point on.
Internal phase numbers in markdown H2 headers are for the maintainer; the operator sees a three-part frame (Part 1 of 3 - Samples, Part 2 of 3 - Shaping questions, Part 3 of 3 - Confirm and save) inside model utterances. Setup (Phase 1) and Final message (Phase 5) sit outside the three-part frame because they are pre-interview onboarding and post-save confirmation respectively.
Cross-skill asymmetry note: this skill has 5 phases. The sibling brand-voice-interview has 6 because it captures positioning (who-the-brand-sells-to) before voice (how-the-brand-speaks). Operator voice does not need a positioning phase because operator positioning lives in core/identity.md (captured by founder-os-setup). Do not collapse positioning into voice when editing the sibling skill.
<Instruction-gate>
Do not generate `core/voice-profile.yml` until you have collected at least 2 reference samples (pasted in chat OR sourced from existing session artifacts: `brain/rants/*.md`, `context/decisions.md`, `brain/log.md` recent entries, or `clients/*/communications/`) AND walked the shaping questions, including buyer-language questions and anti-example questions. Samples are the ground truth. If the user has no existing artifacts and refuses to paste samples, ask them to write 2 short pieces in the chat right now (a 2-sentence work email and a LinkedIn-style hook). Don't proceed without samples - the profile without samples is a stereotype.Do not invent answers. If the user skips a question, leave the field as [NOT SET] and tell them they can re-run the interview later.
</Instruction-gate>
Phase map - load the reference when you reach the phase
Run the phases in order. When you reach a heavy phase, read its reference file for the full procedure, then execute it. Load only the phase you are in.
| Phase | What it does | Load |
| 1 | Setup - operator vs brand, then welcome (inline below) | - |
| 2 | Samples - scan for existing artifacts, paste flow (the ground truth) | references/samples-and-questions.md |
| 3 | Shaping questions - Q1 to Q12 including anti-examples | references/samples-and-questions.md |
| 4 | Confirm and save - the recap block + core/voice-profile.yml output | references/confirm-and-save.md |
| 5 | Final message (inline below) | - |
Phase 1 - Setup (operator vs brand, then welcome)
Say exactly:
Quick check. There are two voice layers in this OS:
- Your personal voice - how YOU write as a person. Personal LinkedIn, your CV, emails from you, anything where your name is on the output. This is what we are about to capture.
- A brand voice - how a brand you run writes. Customer comms, brand social, ads. Captured separately with
/founder-os:brand-voice-interview.If you are running a founder-led brand the two overlap, but the brand voice should still be captured separately so it survives after employees other than you start writing for the brand.
One thing this is NOT: talking to your OS out loud. If you wanted to speak to it and hear it answer, say "add voice" instead and I will set that up.
Is this for you personally, or for a brand you run? (personal / brand)
If they say "brand", stop and dispatch to brand-voice-interview instead.
If they say "personal" or "both", continue with this skill for personal voice. (If "both", note at the end that they should also run brand-voice-interview separately.)
Then say exactly:
Voice interview. About 10 minutes. The output is your personal voice profile - it will write everything as you from now on: emails, posts, scripts, cover letters, whatever you generate. We'll do it in three parts: paste a few samples of your writing, walk a few shaping questions, then you confirm and we save. Ready?
Wait for confirmation.
Phase 2 - Samples (the ground truth)
Read references/samples-and-questions.md. Scan for existing artifacts first, then run the paste flow. Open with "Part 1 - samples." Collect 3 short pieces (or 2 written in chat if the user refuses to paste).
Phase 3 - Shaping questions (Q1-Q12)
Continue in references/samples-and-questions.md. Walk Q1 to Q12 (including the anti-example questions Q9 to Q12), ONE at a time. Open with "Part 2 of 3 - 12 shaping questions."
Phase 4 - Confirm and save
Read references/confirm-and-save.md. Show the recap block (Part 3 of 3), then on a "yes" write core/voice-profile.yml using the exact structure in that reference.
Phase 5 - Final message
Say exactly:
Voice profile saved to core/voice-profile.yml. From now on, anything I generate for you will use this voice - emails, posts, scripts, cover letters, whatever. If something I write doesn't sound like you, tell me and we'll refine the profile.
One more note: this is your PERSONAL voice. If you run a brand (or multiple brands), they have their own voice that should be captured separately. Run
/founder-os:brand-voice-interviewwhen you are ready. The two layers stay independent and writing skills route to the right one based on what you ask for.
Stop. Do not do anything else.
Re-run behavior
If the user runs this skill and core/voice-profile.yml already has real values (not template placeholders), ask:
A voice profile already exists. Want to start over from scratch, or just update specific fields? (start-over / update)
If update, ask which fields, then walk only those questions.
If start-over, run the full interview.
Rules
- One question at a time. Wait for answer.
- Real users ramble. Extract intent. Don't ask them to be more structured.
- Never tell the user to be more concise. The volume is the thinking.
- Never invent samples. Never invent vocabulary.
- Plain language. The user is not a linguist.
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.