agentleFS
slogsdonslogsdon/loop-and-gate-foundation/skills/profile-interview/SKILL.md

profile-interview

Skill0 starsChanged 4 months agoMITView on GitHub

Bootstrap the user's voice profile and taste profile by interviewing them, one question at a time. Use when vault/Profiles/ is empty or stale, or when the user says "build my voice profile" / "taste interview" / "teach the agent how I write". Produces reference docs other skills read so the agent can write in the user's voice and make on-taste calls.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. bas1l/claude-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.

  • In this repository10 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 5 are distinctive, and 0 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.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask which one to run. Default to voice first (immediately useful), taste second.

    line 26
  2. Distinctive:

    One question at a time. Wait for the answer before the next: Never batch.

    line 30
  3. Distinctive:

    Call out contradictions: If an answer clashes with an earlier one, say so.

    line 34
  4. Distinctive:

    Follow the thread: When something unusual surfaces, dig there before moving on.

    line 35
  5. Distinctive:

    Write the profile as a full reference doc — not a summary.

    line 79

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Profile Interview

One job: extract how the user writes and how they judge, into two reference documents precise enough that a skill with no memory of them could still write in their voice and make a call they'd agree with.

All vault/… paths below resolve against the vault base (see CLAUDE.md → "Resolving the vault path"); on a clone that's the repo's vault/, on a plugin install it's the setup-recorded path. Resolve the base before writing.

Two profiles, run as two separate interviews:

  • Voice → vault/Profiles/voice-profile.md. Writing DNA. Consumed by any skill that produces prose — nothing publishes in a voice the kit hasn't been taught.
  • Taste → vault/Profiles/taste-profile.md. Design and product judgment. Consumed at gates where the agent has to decide what's on-brand, what reads as slop, what's worth building. This is how the agent covers a lens the user doesn't have yet.

Ask which one to run. Default to voice first (immediately useful), taste second.

How to interview

  • One question at a time. Wait for the answer before the next. Never batch.
  • Push on vague answers. "I like it simple" → "Simple how? Show me simple done right and simple done lazy." Extract the truth people can't self-report.
  • Ask for real examples. "Show me a sentence you've written that does this."
  • Call out contradictions. If an answer clashes with an earlier one, say so.
  • Follow the thread. When something unusual surfaces, dig there before moving on. The category list is a map, not a script.
  • Don't accept "I don't know" cheaply. Reframe, approach from another angle.
  • The user can stop anytime. On "that's enough" / "compile it", write the profile from what you have and note which areas are thin. A partial profile beats no profile. The interview is re-runnable to deepen it later.

Depth is the user's call: a quick pass (~15 questions, the essentials) or a full pass (the whole map below, ~80–100 questions). Ask which.

Voice interview — the map

Cover these. Counts are for a full pass, scale down for quick.

  • Beliefs & contrarian takes (15) — what they believe that their field doesn't; hot takes they'd defend; conventional wisdom they think is wrong.
  • Writing mechanics (20) — how they actually write vs. how they think they do; default sentence shapes; how they open and close; punctuation, formatting, line breaks; words they overuse, love, would never use.
  • Aesthetic crimes (15) — what makes them cringe in others' writing; phrases like nails on a chalkboard; what reads as lazy.
  • Voice & personality (15) — humor; serious vs. casual tone; how they handle disagreement; excited vs. skeptical.
  • Structural preferences (15) — how they organize; lists/headers/bullets; transitions; default structures.
  • Hard nos (10) — what they'd never write about; approaches they'd never take; lines they won't cross.
  • Red flags (10) — what makes them instantly distrust a piece; signals someone doesn't know what they're talking about.

Taste interview — the map

  • Design judgment (15) — products/sites/objects they admire and why (push past "clean"); what reads as slop; their reaction to specific this-vs-that pairs you show them.
  • Product judgment (15) — what's worth building vs. over-engineered; how they tell a real problem from a loud one; what "good enough to ship" means.
  • Brand & tone (10) — what on-brand feels like for their work; adjectives they want, adjectives they refuse.
  • Hard nos & red flags (10) — design/product patterns they'd never ship; what makes them distrust a product on sight.

Output

Write the profile as a full reference doc — not a summary. Preserve answers, don't compress them away.

# Voice Profile — <name>     (or Taste Profile — <name>)

## Core Identity
<3 sentences — the only summary section>

## <Category>
**Q: <the question you asked>**
<their answer, kept in their words>
... (every answered question, grouped by category)

## Quick Reference Card
**Always:** <specific patterns to follow, pulled from answers>
**Never:** <specific things to avoid>
**Signature phrases & structures:** <real examples they gave>
**Calibration quotes:** <key lines that capture the tone/judgment>

Then add the one-line index entry in vault/MEMORY.md under ## Knowledge index (or a ## Profiles line if you add one), e.g. - [Voice Profile](Profiles/voice-profile.md) — how I write; read before drafting prose.

Rules

  • The profile is the user's, in the user's words. Don't sand their answers into neutral prose — the specificity is the value.
  • One topic per file: voice and taste stay separate documents.
  • Re-running refines, never wipes: merge new answers into the existing profile, keep git history for the rest.
  • The Quick Reference Card is what other skills read most — keep it sharp and concrete, no vague adjectives.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.