Skip to content
Back to search

Prompt Doctor

Rewrite a rough, vague or over-stuffed request into a prompt the current model or agent can act on, and explain a jargon-heavy agent reply in plain words. Use when the user asks to improve, critique or rewrite a prompt, sends a draft before pasting it into another AI, or says "nutshell this" about a reply.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/winsoftlabs30--prompt-doctor-skill/badge.svg)](/listings/winsoftlabs30--prompt-doctor-skill)

HTML

<a href="/listings/winsoftlabs30--prompt-doctor-skill"><img src="/api/listings/winsoftlabs30--prompt-doctor-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

npx skills add WinSoftLabs30/prompt-doctor-skill

On ClawHub (OpenClaw): `openclaw skills install @winsoftlabs30/prompt-doctor`, or `clawhub install @winsoftlabs30/prompt-doctor` with the ClawHub CLI.

Or copy `SKILL.md` into:

- **OpenCode** — `.opencode/skills/prompt-doctor/` in your project, or `~/.config/opencode/skills/prompt-doctor/` for every project - **Claude Code** — `.claude/skills/prompt-doctor/` - **Other agents** — any that read the `SKILL.md` format (Cursor, Gemini CLI, ...)

OpenChamber runs on OpenCode, so the OpenCode path works there too.

No build step, no dependencies, no API key. Any Node version works for this manual edition.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 104 lines
  1. ---
  2. name: Prompt Doctor
  3. description: Rewrite a rough, vague or over-stuffed request into a prompt the current model or agent can act on, and explain a jargon-heavy agent reply in plain words. Use when the user asks to improve, critique or rewrite a prompt, sends a draft before pasting it into another AI, or says "nutshell this" about a reply.
  4. ---
  5. # Prompt Doctor (manual edition)
  6. This edition is the checklist and the judgment. The paid edition adds the CLI
  7. (`pd "draft"`), the model notes database and the launcher that sends the rewritten
  8. prompt into an agent.
  9. When the user asks to improve, critique or rewrite a prompt, or sends a draft
  10. "before I send this": do **not** answer the underlying task. Improve the prompt.
  11. ## Workflow
  12. 1. Read the draft once. Do not guess the project.
  13. 2. Run the checks below in order.
  14. 3. Reply with at most **three** missing pieces, one line each. No lectures.
  15. 4. Give the rewritten prompt in a copyable code block.
  16. 5. Ask one question: "Send this version, or keep the original?"
  17. If the draft is already good, say so in one line and do not rewrite it.
  18. ## Checks
  19. | Check | What to look for |
  20. | --- | --- |
  21. | Vague goal | "fix it", "make it better", "do whatever" with no concrete object |
  22. | No success criteria | nothing says what "done" looks like |
  23. | Missing context | no file, path, version, code or error text |
  24. | More than one ask | several tasks stacked in one message |
  25. | Contradiction | asks for a change and for nothing to change |
  26. | No output format | does not say diff, table, list, length or shape |
  27. | No constraints | nothing says what must not change |
  28. | Scope creep | "also", "while you are at it", "and maybe" |
  29. | Directionless | "be creative", "make it pop" with no reference |
  30. | Permission loop | "ask me before anything" inside an agent (stalls the work) |
  31. | Error not included | "it does not work" without the actual output |
  32. | Urgency instead of detail | "ASAP", "!!", "right now" replacing information |
  33. | Role-play scaffolding | "you are the world's best..." padding |
  34. | No audience or tone | writing task with no reader named |
  35. | Unbounded effort | "cover everything", "no time limit" |
  36. | Hidden assumption | "obviously", "as usual", "like before" |
  37. | Polite filler | padding that carries no instruction (harmless, not the problem) |
  38. | Kitchen sink | a long list of unrelated wants |
  39. ## Rewrite shape
  40. Use only the sections the draft needs:
  41. ```
  42. Goal
  43. <the user's own wording>
  44. Context
  45. - Project and stack: [...]
  46. - What I already tried: [...]
  47. Done when
  48. - <an observable check>
  49. Constraints
  50. - <what must not change>
  51. Output
  52. - <diff / table / under 120 words / ...>
  53. ```
  54. Never invent a file path, an error, a version or a requirement. Leave a bracketed
  55. placeholder and say how many there are.
  56. ## Per target
  57. - **Chat model** (ChatGPT, Claude, Gemini, GLM): goal first, output format, audience for
  58. writing. Drop role-play.
  59. - **Reasoning model** (Opus, Sol Pro, DeepSeek Pro, Grok reasoning): state the decision you
  60. want. Drop "think step by step" and requests to show all reasoning.
  61. - **Agentic coding** (OpenCode, Claude Code, Cursor, Kimi Code, Devstral): file paths, the
  62. command that reproduces, an observable definition of done, what must not change, and an
  63. approval boundary for destructive actions only. Drop mock code and permission loops.
  64. ## Nutshell mode
  65. Trigger: the user pastes a reply from another agent and says "nutshell this", "explain what
  66. it just said", or "what does this mean".
  67. Do **not** continue the original task. Explain the reply, in exactly this shape:
  68. 1. **In a nutshell** - at most 3 plain sentences. No new jargon; if a term cannot be
  69. avoided, define it in everyday words in the same sentence.
  70. 2. **Analogy** - one everyday comparison (cooking, driving, mail, plumbing). One line.
  71. 3. **What it means for you** - one line: can it be ignored, or must the user act?
  72. 4. **What to do next** - one line, only when action is needed. Omit it otherwise.
  73. Hard rules for this mode: never invent facts about the user's project; never add new
  74. tasks; if the reply is already plain, say so in one line and skip the analogy.
  75. ## Hard rules
  76. - Never weaken the user's intent to make the prompt easier.
  77. - Never add role-play, flattery, or "you are an expert" scaffolding.
  78. - Keep the rewrite under 200 words unless the user asks for a longer brief.
  79. - Do not answer the original task when only asked to improve the prompt.
README.md · README · 109 lines
  1. # Prompt Doctor — make a rough prompt one the model can act on
  2. Free agent skill. Paste a draft, get the missing pieces named and a rewritten prompt you
  3. can send.
  4. You know the pattern: you type "fix the login on mobile", the agent guesses, edits five
  5. files, and calls it done. The prompt was the problem, not the model.
  6. This skill turns your agent into a prompt reviewer. It does not answer the task — it
  7. fixes the ask first.
  8. ## What it does
  9. 1. Reads your draft once and refuses to guess your project.
  10. 2. Runs 18 checks in order: vague goal, no success criteria, missing context, more than
  11. one ask, contradictions, no output format, no constraints, scope creep, directionless
  12. requests, permission loops, missing error text, urgency without detail, role-play
  13. padding, missing audience, unbounded effort, hidden assumptions, filler, kitchen sink.
  14. 3. Names at most **three** missing pieces, one line each.
  15. 4. Returns the rewritten prompt in a copyable block, using only the sections the draft
  16. needs: `Goal`, `Context`, `Done when`, `Constraints`, `Output`.
  17. 5. Asks one question: "Send this version, or keep the original?"
  18. Hard rules baked in: never invent a file path, error, version or requirement; leave a
  19. bracketed placeholder instead; never add role-play or "you are an expert" scaffolding;
  20. never answer the task when only asked to improve the prompt.
  21. It adapts to three targets: **chat** models, **reasoning** models (drop "think step by
  22. step"), and **agentic coding** tools (paths, reproduction command, definition of done,
  23. approval boundary for destructive actions only).
  24. ## Install
  25. ```
  26. npx skills add WinSoftLabs30/prompt-doctor-skill
  27. ```
  28. On ClawHub (OpenClaw): `openclaw skills install @winsoftlabs30/prompt-doctor`, or
  29. `clawhub install @winsoftlabs30/prompt-doctor` with the ClawHub CLI.
  30. Or copy `SKILL.md` into:
  31. - **OpenCode** — `.opencode/skills/prompt-doctor/` in your project, or
  32. `~/.config/opencode/skills/prompt-doctor/` for every project
  33. - **Claude Code** — `.claude/skills/prompt-doctor/`
  34. - **Other agents** — any that read the `SKILL.md` format (Cursor, Gemini CLI, ...)
  35. OpenChamber runs on OpenCode, so the OpenCode path works there too.
  36. No build step, no dependencies, no API key. Any Node version works for this manual
  37. edition.
  38. ## Use
  39. Send your agent a draft, or ask:
  40. - "improve this prompt before I send it"
  41. - "why is this prompt bad?"
  42. - "rewrite this so the agent stops guessing"
  43. The skill loads and follows the workflow. It works on any draft, not just coding ones —
  44. emails, briefs, research questions, agent instructions.
  45. ## Nutshell mode
  46. Paste an agent's reply and say **"nutshell this"**. The skill stops working on the task and
  47. explains the reply in at most 3 plain sentences, adds one everyday analogy, says in one
  48. line whether you must act, and one line on what to do next if so. No new jargon, no new
  49. tasks, no invented facts.
  50. ## Try it without installing
  51. Paste this into a ChatGPT Project, Claude Project or Gemini Gem once, and it reviews
  52. every draft you send after that:
  53. ```
  54. Before you answer any request I send, check it for: a concrete goal, success criteria,
  55. context, one task, contradictions, output format, constraints, scope creep, direction,
  56. permission loops, the actual error, urgency without detail, role-play padding, audience,
  57. effort bounds, hidden assumptions, filler, kitchen sink.
  58. If two or more fail: list the missing pieces (max 3, one line each), give me a rewritten
  59. prompt in a code block, then ask "Send the rewritten prompt, or answer the original?".
  60. If nothing important fails, answer normally. Never invent details about my project.
  61. When I paste an agent's reply and say "nutshell": explain it in at most 3 plain sentences,
  62. no new jargon, add one everyday analogy, and say in one line whether I need to act.
  63. ```
  64. ## Manual edition
  65. This repository is the checklist and the judgment. The paid edition adds the CLI
  66. (`pd "draft"` — score 0-100, issues by severity, rewritten prompt, `--json`, `--copy`),
  67. the verified per-model notes database, and the launcher that hands the rewritten prompt
  68. straight to OpenCode or Claude Code.
  69. ## Honest limits
  70. - Rules, not a model. It catches structure, not meaning: a well-structured prompt about
  71. the wrong thing still passes.
  72. - It cannot see your repository or your conversation history. You still supply the facts
  73. it leaves in brackets.
  74. - Adding detail is on you. The skill names what is missing; it never invents the missing
  75. fact.
  76. ## License
  77. MIT-0 — see [LICENSE](LICENSE). Use it, modify it, redistribute it, no attribution
  78. needed. (ClawHub requires MIT-0, so both free channels use one license.)

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.