Home/Catalog/Clarity
Clarity — is it understood?
Brief 17

Copy & Voice Audit

Extract every user-facing string; hunt jargon, vague CTAs, and unhelpful errors; rewrite the worst ten before/after.

COPY.md4 phases7 lenses~2.5k chars

Free & open · no signup · read-only — it ends by asking · nothing leaves your machine

What it does

You are working inside this repo. Mission: extract the words users actually see — buttons, errors, empty states, emails — and audit them for clarity, consistency, and helpfulness. Then rewrite the worst offenders.

Read-only pass. Your only write is the report file.

What it looks for

7 lenses it looks through

Phase 2 sweeps the codebase through every one of these, citing file and line for each finding.

1

Jargon leaks

internal or technical terms surfacing in UI ("entity", "payload", table names)

2

Vague actions

"Submit", "OK", "Continue" where the button should say what happens ("Save changes", "Send invite")

3

Unhelpful errors

messages stating failure without the next step; blame-the-user phrasing

4

Tone drift

formal here, jokey there; exclamation inflation

5

Terminology splits

one concept, multiple names across screens (project/workspace/board)

6

Empty states

screens that describe emptiness instead of inviting the first action

7

Destructive ambiguity

delete/cancel confirmations where the safe choice isn't obvious

The deliverable

What lands in your repo

One structured report at the repo root — or in reports/, if you keep one — the same shape every time, ready for a teammate — or the optional Studio — to act on.

COPY.md
1
Voice snapshot
the tone the strings imply today, in three sentences
2
Worst 10, rewritten
each: location · before · after · why
3
Terminology glossary
concept → the one name, plus the names to retire
4
Voice rules
5 bullets future copy must follow
5
String hygiene
centralization fixes if strings are scattered
One file. Evidence-backed. It ends by asking before touching anything.
How it works

What it makes the agent do

Every Goal Prompt follows the same four steps, so results are consistent and repeatable — no matter which one you run.

1
Phase 1

Extract the strings

Pull user-facing text from components, templates, locale files, email templates, and validation messages.

2
Phase 2

Audit through 7 lenses

Jargon leaks — internal or technical terms surfacing in UI ("entity", "payload", table names).

3
Phase 3

Curate

Pick the 10 worst strings by user impact (frequency × confusion).

4
Phase 4

Report

Create COPY.md at repo root.

Get started

Three ways to run this Goal Prompt

01 · COPY

Paste it in

Copy it and paste it into your agent inside the repo you want checked.

02 · INSTALL

As a slash command

Install the goal plugin once — two commands — then just type /goal:copy-voice-audit.

1/plugin marketplace add GhostlyGawd/goal-prompts
2/plugin install goal@goal-prompts

Or install only this Goal Prompt as /goal-copy-voice-audit:

curl -fsSL https://goal-prompts.vercel.app/install | BRIEF=17 sh
03 · AGENT

From an agent (MCP)

Let an agent fetch it mid-conversation, or pull the raw Goal Prompt by URL.

https://goal-prompts.vercel.app/raw/17.md
Transparency

The exact prompt

Nothing hidden — this is the whole Goal Prompt, verbatim. Read it in a minute, edit it, or copy it as-is.

Read the full brief (2,488 characters)
# Goal: Copy & Voice Audit

You are working inside this repo. Mission: extract the words users actually see — buttons, errors, empty states, emails — and audit them for clarity, consistency, and helpfulness. Then rewrite the worst offenders.

Read-only pass. Your only write is the report file.

## Phase 1 — Extract the strings
- Pull user-facing text from components, templates, locale files, email templates, and validation messages.
- Note where strings live: centralized, scattered, hardcoded?
- From the strings alone, describe the voice this product currently has — it has one, chosen or not.

## Phase 2 — Audit through 7 lenses
1. **Jargon leaks** — internal or technical terms surfacing in UI ("entity", "payload", table names)
2. **Vague actions** — "Submit", "OK", "Continue" where the button should say what happens ("Save changes", "Send invite")
3. **Unhelpful errors** — messages stating failure without the next step; blame-the-user phrasing
4. **Tone drift** — formal here, jokey there; exclamation inflation
5. **Terminology splits** — one concept, multiple names across screens (project/workspace/board)
6. **Empty states** — screens that describe emptiness instead of inviting the first action
7. **Destructive ambiguity** — delete/cancel confirmations where the safe choice isn't obvious

## Phase 3 — Curate
- Pick the 10 worst strings by user impact (frequency × confusion)
- Build the glossary: one name per concept, chosen from what's already most used

## Phase 4 — Report
Create `COPY.md` at repo root:
1. **Voice snapshot** — the tone the strings imply today, in three sentences
2. **Worst 10, rewritten** — each: location · before · after · why
3. **Terminology glossary** — concept → the one name, plus the names to retire
4. **Voice rules** — 5 bullets future copy must follow
5. **String hygiene** — centralization fixes if strings are scattered

Start the report with today's date. If `COPY.md` already exists from a previous run, read it first and lead with what changed since.

## Rules
- Every rewrite must be usable verbatim — real copy, not direction
- Buttons say what they do; errors say what to do next
- No user-facing copy in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
- If a `reports/` directory exists at the repo root, write the report there instead of the root.
- Before asking, present the top findings as a ranked list in plain words
- Report only — end by asking which rewrites to apply

House rules for this Goal Prompt

  • Every rewrite must be usable verbatim — real copy, not direction
  • Buttons say what they do; errors say what to do next
  • No user-facing copy in this repo? Say so in a one-paragraph null report and stop — a null result is a valid finding.
  • If a reports/ directory exists at the repo root, write the report there instead of the root.
  • Before asking, present the top findings as a ranked list in plain words
  • Report only — end by asking which rewrites to apply
Keep exploring

More Clarity briefs

Runs inside these playbooks — curated sequences you can launch with one paste: