Line editor that returns tracked-style edits with reasons
Edits a draft for clarity, concision and tone while keeping the author's voice, and shows every change as before/after with a one-line reason so you can accept or reject each one.
Prompt in Writing by AISkillHub Library
What it does
Most "improve my writing" prompts hand back a rewritten draft, and you then have to diff it by eye to work out what changed and whether the changes were right. This prompt works like a careful human line editor. It keeps your structure and voice, makes the smallest edits that fix real problems, and lists each edit as a numbered before/after pair with the reason (wordiness, ambiguity, passive voice hiding the actor, tone mismatch, inconsistent terms, and so on).
**Use it when** you have a draft that is basically right (a blog post, an internal update, a report section, a cover letter) and you want it tighter without it starting to sound like someone else wrote it.
**What you get** - A short diagnosis: the two or three patterns that cost the most clarity. - A numbered edit list, each with the original text, the suggested text, and the reason. - The fully edited draft with all suggestions applied. - A list of questions where the meaning was unclear and the editor chose not to guess.
**What it will not do** - It will not add facts, statistics, quotes or claims that are not in your draft. If an argument needs evidence, it flags the gap instead of inventing support. - It will not restructure the piece or change your conclusions. For structural problems, use a developmental edit first. - It will not "fix" deliberate style choices you list in the style notes variable.
**Example** Input sentence: "It was decided by the team that the launch would be pushed in order to allow for additional testing to be performed." Edit: "The team delayed the launch to allow more testing." Reason: passive voice hid who decided; "in order to" and "to be performed" added length without meaning.
How to use it
## How to use
1. Open `prompts/line-editor-tracked-changes.md` and copy everything below the title line. 2. Replace each `{{variable}}` with your own content. Paste long material (notes, transcripts, tables) directly in place of the variable; keep the surrounding tags so the model can tell inputs apart. 3. Start a new chat in ChatGPT, Claude or Gemini and paste the filled prompt as your first message. 4. Read the output against the quality checks at the end of the prompt before you use or send it.
## Variables
- `{{draft}}`: the full text you want edited. - `{{audience}}`: who will read it, e.g. `executive team`, `first-time customers`, `hiring manager at a fintech`. - `{{goal}}`: what the piece must achieve, e.g. `get approval for budget`, `explain the outage`. - `{{style_notes}}`: house style or choices to keep, e.g. `British spelling, keep first person, Oxford comma`. Optional. - `{{edit_depth}}`: `light` (errors and obvious clutter only), `standard`, or `firm` (cut harder, up to about 25% shorter).
If you do not have a value for an optional variable, write `none` rather than deleting it, so the prompt still reads correctly.
## Reusing it
To reuse it often, save the prompt (with the variables left in) as the instructions of a Claude Project, a custom GPT or a Gemini Gem, then paste only the inputs in each new chat.
## Tips
- Run it at `light` first on important documents; you can always ask for a `firm` pass afterwards. - Paste back only the edit numbers you reject and ask for a revised draft without them.
Files
Save each file at the path shown, relative to your project root.