本文へ移動
cccskills
無料GitHub で公開

interpreter-design

Design or revise Interpreter UI for non-developers with an ultra-minimal, utilitarian style inspired by the OpenAI website and ChatGPT, not the developer platform. Use when simplifying settings, onboarding, chat, forms, empty states, or helper copy; when removing extra containers, labels, or visual noise; when replacing technical wording with plain language; when clarifying hierarchy and making the main user action obvious; when making every visible label and status understandable to an average HR admin or other non-technical user; when hiding implementation details such as logs, IDs, endpoints, paths, or protocol names; or when a screen must respect the app's existing CSS variables and light/dark theme behavior.

インストール方法を見る

含まれるファイル(10)

  • SKILL.md6.8 KB
  • agents/openai.yaml331 B
  • assets/openai/chatgpt-chat-ui.jpg21.9 KB
  • assets/openai/chatgpt-download-page.jpg54.0 KB
  • assets/openai/chatgpt-home.jpg12.3 KB
  • assets/openai/openai-contact-sales.jpg20.3 KB
  • assets/openai/openai-home-prompt.jpg7.3 KB
  • assets/openai/openai-search-overlay.jpg5.5 KB
  • references/openai-screen-index.md2.1 KB
  • references/openai-style-guide.md2.0 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Interpreter Design

Design for non-developers. Keep the UI quiet, direct, and easy to scan.

Workflow

  1. Read references/openai-style-guide.md.
  2. Skim references/openai-screen-index.md.
  3. Open the bundled screenshots in assets/openai/ if visual calibration is useful.
  4. Before changing layout, write this brief to yourself:
    • The user is here to do:
    • The main action is:
    • The user can ignore:
  5. Start from the least UI that still explains the task and lets the person act.
  6. Run the final reduction pass before finishing.

Rules

  • Every visible label, status, and explanation must make sense to an average HR admin on first read.
  • Design around the user's task, not the system's internals.
  • Decide what the user most likely needs to do next before choosing a layout.
  • Keep the primary action obvious. Keep secondary actions quiet.
  • If there is one clear next step, give it clear visual priority.
  • If two controls do different jobs, do not style them as equally important by accident.
  • Do not place a low-priority note opposite a high-priority action in the same row.
  • For install, onboarding, and help surfaces, assume one centered instruction block is enough until proven otherwise.
  • Use sentence case. Do not use all-caps titles, eyebrow labels, or loud helper text.
  • Prefer spacing, alignment, and thin dividers over cards, nested panels, or heavy borders.
  • Use as few containers as possible. Add a box only when it clearly improves comprehension.
  • Write short labels and short helper text. Cut filler, hype, and repeated explanation.
  • Minimal does not mean cramped. Leave enough room for the main action and the main explanation to breathe.
  • Remove developer language unless the feature is explicitly for developers.
  • Do not expose technical implementation detail in the visible UI. Hide logs, file paths, IDs, endpoints, protocol names, process states, and similar internals.
  • If the product needs internal diagnostics, keep them out of the main interface for non-developers.
  • Do not design like the OpenAI developer platform. Use the OpenAI website and ChatGPT product surfaces only.
  • Respect the app's existing CSS variables, tokens, and theme system. Do not hardcode a separate light or dark palette when app vars already cover it.
  • Make light mode and dark mode both work by leaning on existing variables, not one-off overrides.

Layout

  • Start with one clean column unless the task truly needs more.
  • Use side-by-side layouts only when both sides are comparable in importance and type.
  • Do not use a hero layout for a settings surface unless the task truly needs one.
  • If one heading, one action line, and one quiet fallback line are enough, stop there.
  • Keep headings short.
  • Avoid narrow columns packed with long text.
  • If a row feels pinched, stack it or widen it.
  • Group related controls by proximity first, separators second, containers last.
  • Let whitespace do most of the structure work.
  • Keep explanations close to the control they clarify.
  • Avoid dashboard density, enterprise framing, and control-panel chrome.

Hierarchy

  • The screen should answer, in order:
    1. What is this?
    2. What should I do?
    3. What else might I need to know?
  • Put the likely next action where the eye lands first.
  • Keep status and diagnostics below the action unless they are required to choose the action.
  • Notes, caveats, and "this is optional" copy are footnotes. They should not compete with controls.
  • Metrics should support a decision, not exist as decoration.
  • If a section has a single important action, make the section feel built around that action.
  • If the page feels visually balanced but the next step is still unclear, the hierarchy is wrong.
  • If the next step is clear but the layout feels cramped, lopsided, or top-heavy, the balance is wrong.

Balance

  • Aim for calm visual balance, not perfect symmetry.
  • Do not let one side of a row carry all the weight while the other side carries a footnote or tiny fragment.
  • Keep text blocks, actions, and status areas proportionate.
  • If a button looks crowded by nearby copy, give it room or change the layout.
  • If the top of the screen feels squeezed, reduce copy, widen the content, or stack the elements.
  • A simple screen should feel settled and spacious, not compressed.
  • Ask: "Does this feel calm and balanced at a glance?"

Copy

  • Use plain words a non-developer understands on first read.
  • Write as if the reader works in HR, recruiting, operations, or support, not engineering.
  • Explain only what helps the person decide or act.
  • Replace technical terms with direct labels.
  • Do not make the person learn system architecture to use the feature.
  • If a line does not help someone complete the task, delete it.

Action audit

  • List the actions the user can take on this screen.
  • Mark one as primary, if one truly is primary.
  • Mark the rest as secondary, supportive, or removable.
  • Check that the visual weight matches that ranking.
  • If you cannot name the primary action, the screen is probably trying to do too much.

Final pass

  • Remove one more container.
  • Try replacing the whole screen with one centered instruction block. If the task still works, keep the simpler version.
  • Shorten every heading.
  • Cut any line that repeats nearby context.
  • Replace technical terms with plain language.
  • Remove any visible information an average HR admin would not immediately understand.
  • Remove any logs, IDs, endpoints, file paths, protocol names, or implementation detail from the main UI.
  • Check that the first thing the user sees is the thing they most likely need to do.
  • Check that helper text, notes, and statuses do not compete with the main action.
  • Ask: "What is this person here to do, and did I make that easy?"
  • Ask: "Does the composition feel balanced, calm, and easy to scan?"
  • Check that the screen still reads clearly in both light and dark mode using existing app vars.
  • Ask: "Would a non-developer understand this immediately?" If not, simplify again.

References

  • Use references/openai-style-guide.md for the distilled rules.
  • Use references/openai-screen-index.md for the bundled OpenAI screenshots and what to learn from each.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Build a period-end accrual schedule by computing each accrual, citing support, and drafting journal entries for controller approval. Use during month-end close; this drafts entries only and must not post them.

日本語の概要は準備中です。原文の説明を表示しています。

openinterpreter/interpreter-workstation232026年10月11日 更新

audit-xls

無料

Audit a spreadsheet for formula accuracy, errors, and common financial-model mistakes. Use for selected ranges, single sheets, or whole-workbook model checks including balance-sheet balance, cash tie-out, roll-forwards, and logic sanity.

日本語の概要は準備中です。原文の説明を表示しています。

openinterpreter/interpreter-workstation232026年10月11日 更新

Use this skill when the user needs advanced Playwright control of an already running browser session through Interpreter's app-managed browser bridge and the `builtin-js-repl` `js_repl` tool, after simple browser page work cannot be handled by the unified `builtin-interpreter` browser page tools.

日本語の概要は準備中です。原文の説明を表示しています。

openinterpreter/interpreter-workstation232026年10月11日 更新

Drive a native desktop app through Interpreter's builtin-cua-driver. Use get_app_state, click/type/scroll/drag, and verify by calling get_app_state again when the user asks to operate a real desktop app, browser chrome, native dialog, menu, secure prompt, file chooser, or hidden/background window.

日本語の概要は準備中です。原文の説明を表示しています。

openinterpreter/interpreter-workstation232026年10月11日 更新

doc

無料

Create, inspect, and edit Word documents (`.docx`) with local code-execution tools. Use for document drafting, structured edits, tables, comments, and layout-sensitive Word deliverables.

日本語の概要は準備中です。原文の説明を表示しています。

openinterpreter/interpreter-workstation232026年10月11日 更新

Use for AI image, video, audio, or 3D generation/editing. Prefer configured media tools when present, estimate cost before running a paid model, and keep provider credentials in the user's environment or provider configuration.

日本語の概要は準備中です。原文の説明を表示しています。

openinterpreter/interpreter-workstation232026年10月11日 更新

openinterpreter のスキルをすべて見る

このスキルの問題を報告する