Дизайнерское перо в стиле Cursor для Claude Code. Кликайте, выделяйте рамкой или рисуйте на любой веб-странице в окне Chromium, и Claude получит точный исходный файл, номер строки, CSS-селектор и скриншот — готовые к редактированию и проверке.
Три режима визуального поиска исходного кода:
Режим | Взаимодействие | Claude получает |
element | Наведение + клик по элементу |
|
area | Перетаскивание рамки |
|
draw | Свободное рисование красным пером, Enter для завершения |
|
Все скриншоты сохраняются как PNG-файлы в /tmp и возвращаются в виде путей — ваш MCP-клиент никогда не превысит лимит контекста из-за base64.
Разрешение исходного кода React работает в режиме разработки Next.js через свойство fiber _debugSource (добавляется с помощью @babel/plugin-transform-react-jsx-source). В производственных сборках это свойство удаляется; см. Production source mapping ниже.
Related MCP server: software-design-mermaid-mcp
You: "Make this button rounder"
Claude: [designer_open http://localhost:3000/dashboard]
Claude: [designer_pick mode=element]
You: *click the button*
Claude: → source: Button.tsx:42
Claude: [Edit Button.tsx add rounded-full]
Claude: [designer_screenshot selector=#cta-btn] ← after screenshot for verificationПредварительные требования: Node 18+, Claude Code, работающая macOS/Linux (Playwright Chromium).
git clone https://github.com/YOUR_USER/designer-mcp.git
cd designer-mcp
npm install
npx playwright install chromium # one-time browser downloadЗарегистрируйте MCP в Claude Code (user-scope = доступно в каждой сессии):
claude mcp add --scope user designer-mcp node "$(pwd)/index.js"Установите навык Claude, чтобы будущие сессии знали об этом рабочем процессе:
mkdir -p ~/.claude/skills/designer
cp SKILL.md ~/.claude/skills/designer/SKILL.mdПерезапустите Claude Code. Вы должны увидеть инструменты designer_* и навык designer: в вашей сессии.
Запустите сервер разработки Next.js (для сопоставления исходного кода):
cd your-nextjs-app && npm run devЗатем в Claude Code:
"Открой http://localhost:3000/settings в дизайнере и позволь мне выбрать заголовок."
Claude вызовет designer_open(...), затем designer_pick({ mode: "element" }). Окно Chromium выйдет на передний план, ваш курсор превратится в перекрестие, вы кликнете по заголовку. Claude получит source.fileName + lineNumber и сможет сразу приступить к редактированию.
Один элемент — используйте element
Несколько связанных элементов в одной области — используйте area (перетащите рамку; возвращает каждый элемент, чей центр попадает внутрь)
Аннотирование / визуальное объяснение — используйте draw (красное перо, Enter для завершения, Esc для отмены)
_debugSource доступен только в режиме разработки. Чтобы использовать инструмент выбора в производственной сборке, включите карты исходного кода (source maps) в next.config.js:
module.exports = {
productionBrowserSourceMaps: true,
// ...
};В настоящее время инструмент выбора возвращает source: null в продакшене; будущая версия будет разрешать селектор через развернутую карту исходного кода. Пул-реквесты приветствуются.
Все инструменты доступны через MCP; Claude Code видит их как mcp__designer-mcp__*.
designer_open(url: string)Запускает или повторно использует запущенный экземпляр Chromium и переходит по адресу. Выводит окно на передний план в macOS с помощью bringToFront() + AppleScript.
designer_pick({ mode?: "element" | "area" | "draw" })Активирует оверлей выбора. Возвращает результат, когда пользователь завершает взаимодействие (или Esc для отмены, или по тайм-ауту 180 секунд).
designer_screenshot({ selector?: string })PNG страницы или конкретного элемента. Возвращает { path, bytes }.
designer_close()Закрывает браузер и освобождает ресурсы Playwright.
Запускается управляемый Playwright экземпляр Chromium. Один экземпляр на процесс.
designer_pick внедряет небольшой оверлей на чистом JS (picker.js) на страницу. Оверлей:
режим element — отслеживает mousemove/click, выделяет цель наведения синим цветом, разрешает уникальный CSS-селектор, проходит по цепочке React fiber для поиска _debugSource, возвращает результат в MCP.
режим area — выделение рамкой; при отпускании кнопки мыши собираются все элементы, чей центр попадает внутрь рамки (дедупликация по селектору).
режим draw — оверлей холста на весь экран; захватывает штрихи как массивы точек; Enter завершает работу.
Сервер опрашивает window.__designerResult каждые 200 мс в течение до 180 секунд.
По завершении соответствующий скриншот (элемент / фрагмент области / весь экран) сохраняется в /tmp и возвращается путь.
Пул-реквесты приветствуются, особенно в следующих областях:
Разрешение карт исходного кода в продакшене
Инструмент выбора для Kestrel/React Native (сейчас только веб)
Накопление нескольких элементов в режиме element (Cmd-клик для добавления)
Интеграция "показать в редакторе" для VS Code
MIT
Close the designer browser and release resources.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It mentions 'release resources', which hints at cleanup behavior, but does not disclose critical details like whether this is destructive (e.g., closes without saving), requires specific permissions, or has side effects. For a tool with no annotations, this leaves significant behavioral gaps.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence that directly states the tool's purpose with no wasted words. It is appropriately sized and front-loaded, making it easy to understand immediately without unnecessary elaboration.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity (0 parameters, no output schema, no annotations), the description is adequate but incomplete. It covers the basic action but lacks details on behavioral aspects like what happens to unsaved work or error conditions. For a tool that likely interacts with a browser, more context would be helpful despite the low complexity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has 0 parameters with 100% schema description coverage, so the schema fully documents the lack of inputs. The description does not add parameter details, which is unnecessary here. Baseline is 4 for 0 parameters, as no additional parameter semantics are needed beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the specific action ('Close') and resource ('the designer browser'), distinguishing it from sibling tools like designer_open (open), designer_pick (pick), and designer_screenshot (capture screenshot). It provides a complete verb+resource combination that is unambiguous.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage when the designer browser is open and resources need releasing, but it does not explicitly state when to use this tool versus alternatives or any prerequisites. It lacks explicit guidance on when-not-to-use or named alternatives, leaving usage context somewhat implied rather than clearly defined.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Open a URL in the designer's headed Chromium (launches it if not running). Use this before designer_pick.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | URL to navigate to |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It discloses key behavioral traits: it opens a URL, launches Chromium if not running, and is intended as a preparatory step. However, it lacks details on error handling, timeouts, or what happens if the URL is invalid, which are important for a tool that interacts with a browser.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise and front-loaded, with two sentences that efficiently convey the action, conditional behavior, and usage context. Every sentence earns its place by providing essential information without redundancy or fluff.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's low complexity (1 parameter, no output schema, no annotations), the description is mostly complete. It covers purpose, usage, and key behavior. However, it could improve by mentioning potential side effects (e.g., browser window focus) or error scenarios, slightly reducing completeness for a tool that launches external processes.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, with the schema fully documenting the 'url' parameter. The description does not add any meaning beyond the schema (e.g., URL format examples or constraints), so it meets the baseline of 3 where the schema does the heavy lifting without extra value from the description.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the specific action ('Open a URL') and the target resource ('in the designer's headed Chromium'), including the conditional behavior ('launches it if not running'). It distinguishes from siblings by mentioning designer_pick as a subsequent step, showing awareness of the tool's role in the workflow.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It explicitly provides when-to-use guidance by stating 'Use this before designer_pick,' establishing a clear sequence in the workflow. This directly addresses when to use this tool versus alternatives (like designer_screenshot or designer_close) by positioning it as a prerequisite step.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Activate the picker in the designer browser. Three modes: element — user clicks one element; returns { selector, tag, classes, text, html, rect, source, screenshot_path } area — user drags a marquee; returns { rect, elements: [{selector, source, rect, ...}], screenshot_path } draw — user ink-annotates with a red pen, Enter to finish; returns { strokes, viewport, screenshot_path (strokes only), viewport_screenshot_path (full view with drawings) } Esc cancels in any mode. screenshot_path / viewport_screenshot_path point to PNGs in /tmp; open with the Read tool.
| Name | Required | Description | Default |
|---|---|---|---|
| mode | No | element (default) = click one, area = drag marquee, draw = freeform pen |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes the interactive nature of the tool (user clicks/drags/annotates), cancellation behavior, and output file handling (PNGs in /tmp). However, it doesn't mention potential side effects like browser focus changes or performance considerations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is efficiently structured with bullet-like formatting for the three modes, each clearly explaining the user interaction and return values. Every sentence adds essential information about functionality, cancellation, or output handling with zero wasted text.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool with no annotations and no output schema, the description provides comprehensive context about the interactive process, return data structures, and file outputs. The only minor gap is lack of explicit mention about whether this tool requires specific browser state or permissions.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema has 100% description coverage with clear enum values and descriptions. The description adds significant value by detailing what each mode returns (specific data structures like selector, rect, strokes, etc.) and operational differences between modes, going well beyond the schema's basic mode definitions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Activate the picker in the designer browser' with three specific modes (element, area, draw). It distinguishes from siblings like designer_close, designer_open, and designer_screenshot by focusing on interactive element/area selection and annotation rather than basic browser operations or screenshot capture.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit guidance on when to use each mode: element for clicking one element, area for dragging a marquee, and draw for freeform pen annotation. It also specifies 'Esc cancels in any mode' and mentions using the Read tool to open resulting PNGs, giving clear operational context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Screenshot the current page or a specific element selector. Returns { path, bytes } — a filesystem path to a PNG in /tmp that you can Read with the Read tool.
| Name | Required | Description | Default |
|---|---|---|---|
| selector | No | Optional CSS selector |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden and effectively discloses key behaviors: it returns a filesystem path to a PNG in /tmp, specifies the output format ({ path, bytes }), and mentions a follow-up action (Read tool). However, it lacks details on potential errors or limitations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is front-loaded with the core purpose, followed by essential details on output and usage, with every sentence earning its place and no wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity (screenshot functionality with one optional parameter) and no output schema, the description is mostly complete, covering purpose, output, and a follow-up action, though it could include more on error handling or constraints.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema description coverage is 100%, so the baseline is 3. The description adds minimal value beyond the schema by implying the selector is optional and used for targeting elements, but does not provide additional syntax or format details.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Screenshot') and resource ('the current page or a specific element selector'), distinguishing it from sibling tools like designer_close, designer_open, and designer_pick by specifying its unique screenshot functionality.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It provides clear context for usage by specifying 'the current page or a specific element selector' and mentions an alternative action ('Read with the Read tool'), but does not explicitly state when not to use it or compare directly to sibling tools.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0designer_closedesigner_opendesigner_pickdesigner_screenshotScored across 4 tools
Each tool has a clearly distinct purpose with no overlap: open launches the browser, pick activates the picker with specific modes, screenshot captures images, and close terminates the session. The descriptions clearly differentiate their functions, eliminating any ambiguity in tool selection.
All tool names follow a consistent 'designer_' prefix with descriptive action suffixes (open, pick, screenshot, close), using snake_case uniformly. This predictable pattern makes the tool set easy to navigate and understand at a glance.
With 4 tools, this server is well-scoped for its purpose of browser-based design interactions. Each tool earns its place by covering essential operations: launching, picking elements, capturing screenshots, and cleaning up, without being overly sparse or bloated.
The tool set provides complete lifecycle coverage for the designer domain: it supports opening the browser, interactive element selection, screenshot capture, and proper resource closure. There are no obvious gaps, as all core workflows from initiation to termination are addressed effectively.
Build, clone & publish websites by chatting with Claude. Live in seconds, custom domains + SSL.
Live SEO workflow tools for Claude Code, Codex, and AI agents.
Comment on AI-generated webpages; feedback flows back to your coding agent. Free, MIT, local-first.
Agent-Native design tool - create and edit visual designs with agent assistance
