El servidor de referencia del Model Context Protocol para trabajo frontend. Una instalación le da a cualquier agente de programación — Claude Desktop, Cursor, VS Code, Zed, Continue — ojos y manos reales sobre tu interfaz:
axe_audit — ejecuta el motor de reglas real axe-core contra HTML sin procesar (vía jsdom) o una URL en vivo (vía Playwright). Devuelve violaciones agrupadas por impacto WCAG con enlaces de corrección.
page_screenshot — PNGs de Chromium headless de cualquier URL o selector CSS, con controles de viewport / DPR / esquema de color / waitForSelector.
bundle_budget_check — recorre un directorio dist/, calcula tamaños gzip (+ brotli opcional), aplica un presupuesto en KB global o por entrada. JSON de aprobado/fallo listo para CI.
design_token_diff — diferencia estructural de dos archivos de tokens W3C DTCG / Style Dictionary. Informa de tokens añadidos / eliminados / cambiados con notas que reconocen $type.
storybook_story_run — carga una historia individual de Storybook en Chromium headless vía iframe.html?id=…, toma una captura de pantalla y ejecuta una auditoría axe solo sobre el componente renderizado.
scaffold_react_component — emite un componente React tipado (funcional / forwardRef / as polimórfico) además de pruebas Vitest e historias de Storybook opcionales.
Todas las herramientas devuelven JSON estructurado sobre el que el modelo puede razonar. Sin envoltorios personalizados por editor.
npm i -g @ashios15/mcp-frontend-tools
# Optional — enables page_screenshot, storybook_story_run, and URL-mode axe_audit
npm i -g playwright
npx playwright install chromiumSe requiere Node ≥ 20. Sin Playwright, el servidor aún se inicia; esas tres herramientas devolverán un error claro de "instalar playwright" si se llaman.
Añadir a ~/Library/Application Support/Claude/claude_desktop_config.json (o el equivalente en Windows):
{
"mcpServers": {
"frontend-tools": {
"command": "mcp-frontend-tools"
}
}
}Settings → MCP → Add new server:
{
"frontend-tools": { "command": "mcp-frontend-tools" }
}Añadir a .vscode/mcp.json:
{
"servers": {
"frontend-tools": { "command": "mcp-frontend-tools" }
}
}npx @modelcontextprotocol/inspector mcp-frontend-toolsRelated MCP server: UX/UI Tools for React + Material-UI
axe_audit{
url?: string; // require playwright
html?: string; // jsdom
tags?: string[]; // default: wcag2a/aa + wcag21aa + wcag22aa + best-practice
selector?: string; // URL mode only
timeoutMs?: number; // default 15000
}Devuelve violaciones y reglas incompletas con impact, help, helpUrl y hasta 5 nodos de ejemplo fallidos por regla.
page_screenshot{
url: string;
outPath: string; // absolute; parent dirs auto-created
selector?: string;
fullPage?: boolean;
width?: number; // default 1280
height?: number; // default 800
deviceScaleFactor?: number;// default 2
colorScheme?: "light" | "dark" | "no-preference";
waitForSelector?: string;
timeoutMs?: number;
}bundle_budget_check{
buildDir: string;
budgetKb?: number; // default 250 (gzipped)
perEntryBudgetKb?: Record<string, number>;// key = path substring, longest match wins
ext?: string[]; // default [".js",".mjs",".cjs",".css"]
includeBrotli?: boolean;
}Devuelve cada archivo con tamaños raw / gzip / brotli, su presupuesto aplicado y status: "pass" | "fail".
design_token_diff{
beforePath: string;
afterPath: string;
ignoreKeys?: string[];
}Entiende la forma W3C DTCG ($value, $type). Las claves de metadatos de grupo que comienzan con $ se ignoran. Los cambios de color / dimensión obtienen notes anotadas.
storybook_story_run{
storybookUrl: string; // e.g. http://localhost:6006
storyId: string; // e.g. components-button--primary
screenshotPath?: string;
runAxe?: boolean; // default true
viewport?: { width: number; height: number };
colorScheme?: "light" | "dark" | "no-preference";
timeoutMs?: number;
}Carga ${storybookUrl}/iframe.html?viewMode=story&id=${storyId}, espera a #storybook-root e informa de cualquier error de página + violaciones de axe limitadas a la historia.
scaffold_react_component{
name: string; // PascalCase
outDir: string; // absolute
variant?: "functional" | "forwardRef" | "polymorphic";
withTests?: boolean; // default true
withStory?: boolean; // default true
props?: Array<{ name: string; type: string; required?: boolean; defaultValue?: string }>;
}Los agentes obtienen contratos tipados. Cada herramienta es un JSON Schema validado por Zod que el modelo puede inspeccionar.
Sin alucinaciones locales. Las violaciones de Axe provienen del motor real axe-core, no de una expresión regular. Los tamaños de paquete provienen de la compresión real zlib, no de estimaciones.
Componible. Fallo en bundle_budget_check → pide al agente que lleve el archivo más grande a page_screenshot → alimenta la imagen en un paso de regresión visual. Todo a través de MCP.
Superficie estable. Las herramientas cambian detrás de un aumento de versión; tu .vscode/mcp.json no.
git clone https://github.com/ashios15/mcp-frontend-tools.git
cd mcp-frontend-tools
npm install
npm run test # 3 unit tests (bundle + tokens + scaffold)
npm run build
npm run inspector # opens MCP Inspector against the built server
node scripts/smoke.mjs # quick stdio tools/list checkMIT © ashios15
El servidor de referencia del Model Context Protocol para trabajo frontend. Una instalación le da a cualquier agente de programación — Claude Desktop, Cursor, VS Code, Zed, Continue — ojos y manos reales sobre tu interfaz:
axe_audit — ejecuta el motor de reglas real axe-core contra HTML sin procesar (vía jsdom) o una URL en vivo (vía Playwright). Devuelve violaciones agrupadas por impacto WCAG con enlaces de corrección.
page_screenshot — PNGs de Chromium headless de cualquier URL o selector CSS, con controles de viewport / DPR / esquema de color / waitForSelector.
bundle_budget_check — recorre un directorio dist/, calcula tamaños gzip (+ brotli opcional), aplica un presupuesto en KB global o por entrada. JSON de aprobado/fallo listo para CI.
design_token_diff — diferencia estructural de dos archivos de tokens W3C DTCG / Style Dictionary. Informa de tokens añadidos / eliminados / cambiados con notas que reconocen $type.
storybook_story_run — carga una historia individual de Storybook en Chromium headless vía iframe.html?id=…, toma una captura de pantalla y ejecuta una auditoría axe solo sobre el componente renderizado.
scaffold_react_component — emite un componente React tipado (funcional / forwardRef / as polimórfico) además de pruebas Vitest e historias de Storybook opcionales.
Todas las herramientas devuelven JSON estructurado sobre el que el modelo puede razonar. Sin envoltorios personalizados por editor.
npm i -g @ashishjoshi/mcp-frontend-tools
# Optional — enables page_screenshot, storybook_story_run, and URL-mode axe_audit
npm i -g playwright
npx playwright install chromiumSe requiere Node ≥ 20. Sin Playwright, el servidor aún se inicia; esas tres herramientas devolverán un error claro de "instalar playwright" si se llaman.
Añadir a ~/Library/Application Support/Claude/claude_desktop_config.json (o el equivalente en Windows):
{
"mcpServers": {
"frontend-tools": {
"command": "mcp-frontend-tools"
}
}
}Settings → MCP → Add new server:
{
"frontend-tools": { "command": "mcp-frontend-tools" }
}Añadir a .vscode/mcp.json:
{
"servers": {
"frontend-tools": { "command": "mcp-frontend-tools" }
}
}npx @modelcontextprotocol/inspector mcp-frontend-toolsaxe_audit{
url?: string; // require playwright
html?: string; // jsdom
tags?: string[]; // default: wcag2a/aa + wcag21aa + wcag22aa + best-practice
selector?: string; // URL mode only
timeoutMs?: number; // default 15000
}Devuelve violaciones y reglas incompletas con impact, help, helpUrl y hasta 5 nodos de ejemplo fallidos por regla.
page_screenshot{
url: string;
outPath: string; // absolute; parent dirs auto-created
selector?: string;
fullPage?: boolean;
width?: number; // default 1280
height?: number; // default 800
deviceScaleFactor?: number;// default 2
colorScheme?: "light" | "dark" | "no-preference";
waitForSelector?: string;
timeoutMs?: number;
}bundle_budget_check{
buildDir: string;
budgetKb?: number; // default 250 (gzipped)
perEntryBudgetKb?: Record<string, number>;// key = path substring, longest match wins
ext?: string[]; // default [".js",".mjs",".cjs",".css"]
includeBrotli?: boolean;
}Devuelve cada archivo con tamaños raw / gzip / brotli, su presupuesto aplicado y status: "pass" | "fail".
design_token_diff{
beforePath: string;
afterPath: string;
ignoreKeys?: string[];
}Entiende la forma W3C DTCG ($value, $type). Las claves de metadatos de grupo que comienzan con $ se ignoran. Los cambios de color / dimensión obtienen notes anotadas.
storybook_story_run{
storybookUrl: string; // e.g. http://localhost:6006
storyId: string; // e.g. components-button--primary
screenshotPath?: string;
runAxe?: boolean; // default true
viewport?: { width: number; height: number };
colorScheme?: "light" | "dark" | "no-preference";
timeoutMs?: number;
}Carga ${storybookUrl}/iframe.html?viewMode=story&id=${storyId}, espera a #storybook-root e informa de cualquier error de página + violaciones de axe limitadas a la historia.
scaffold_react_component{
name: string; // PascalCase
outDir: string; // absolute
variant?: "functional" | "forwardRef" | "polymorphic";
withTests?: boolean; // default true
withStory?: boolean; // default true
props?: Array<{ name: string; type: string; required?: boolean; defaultValue?: string }>;
}Los agentes obtienen contratos tipados. Cada herramienta es un JSON-Schema validado por Zod que el modelo puede inspeccionar.
Sin alucinaciones locales. Las violaciones de Axe provienen del motor real axe-core, no de una expresión regular. Los tamaños de paquete provienen de la compresión real zlib, no de estimaciones.
Componible. Fallo en bundle_budget_check → pide al agente que lleve el archivo más grande a page_screenshot → alimenta la imagen en un paso de regresión visual. Todo a través de MCP.
Superficie estable. Las herramientas cambian detrás de un aumento de versión; tu .vscode/mcp.json no.
npm install
npm run test # 3 unit tests (bundle + tokens + scaffold)
npm run build
npm run inspector # opens MCP Inspector against the built server
node scripts/smoke.mjs # quick stdio tools/list checkMIT © Ashish Joshi
Un servidor de Model Context Protocol (MCP) que da a los asistentes de IA (Claude, Copilot, Cursor) acceso a herramientas de desarrollo frontend — creación de componentes, análisis de paquetes, comprobaciones de accesibilidad y guías de diseño responsivo.
Herramienta | Descripción |
| Genera un componente React tipado con pruebas, historias y módulo CSS |
| Escanea un directorio de compilación en busca de JS/CSS sobredimensionado, informa de los hallazgos |
| Comprobaciones estáticas WCAG 2.2 en HTML con sugerencias de corrección |
| Genera patrones de CSS responsivo, consultas de contenedor y Tailwind |
Añadir a ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"frontend-tools": {
"command": "node",
"args": ["/path/to/mcp-frontend-tools/dist/index.js"]
}
}
}Añadir a .vscode/settings.json:
{
"github.copilot.chat.mcpServers": {
"frontend-tools": {
"command": "node",
"args": ["${workspaceFolder}/mcp-frontend-tools/dist/index.js"]
}
}
}"Crea un componente UserProfileCard con props de avatar, nombre y biografía"
La IA llama a scaffold_react_component y obtiene:
UserProfileCard.tsx — Componente tipado con forwardRef
UserProfileCard.test.tsx — Pruebas con Testing Library
UserProfileCard.stories.tsx — Historia de Storybook
UserProfileCard.module.css — Módulo CSS
index.ts — Exportación de barril
"Analiza mi carpeta dist/ en busca de problemas de tamaño de paquete"
La IA llama a analyze_bundle y devuelve un informe en markdown con archivos sobredimensionados, recomendaciones y una tabla resumen.
src/
├── index.ts # MCP server setup (stdio transport)
└── tools/
├── index.ts # Tool definitions + router
├── scaffold-component.ts # React component generator
├── bundle-analyzer.ts # Build output analyzer
├── a11y-checker.ts # Static WCAG checks
└── responsive-guide.ts # Responsive CSS pattern generatornpm install
npm run build
npm run inspector # Test with MCP InspectorMIT
Run the axe-core accessibility ruleset against an HTML string (jsdom) or a live URL (Playwright). Returns violations grouped by impact with fix guidance and helpUrl links.
| Name | Required | Description | Default |
|---|---|---|---|
| url | No | Fully qualified URL to audit. Requires playwright installed. | |
| html | No | Raw HTML string to audit (jsdom, no JS execution). | |
| tags | No | axe rule tags to include. Defaults to ['wcag2a','wcag2aa','wcag21a','wcag21aa','wcag22aa','best-practice']. | |
| selector | No | CSS selector to scope the audit (URL mode only). | |
| timeoutMs | No | Default 15000. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It discloses the audit scope (axe-core), return format (grouped violations with guidance/links), and that URL mode uses Playwright. No contradictions.
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?
Two sentences, no fluff. First sentence states core action and modes, second describes output format. Information density is high.
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?
No output schema, but description sufficiently hints at return structure (violations grouped by impact, fix guidance, helpUrl). Parameters well described, tool behavior clear. Complete for the complexity level.
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 has 100% description coverage. Description adds context beyond schema: clarifies dual mode (url/html), defaults for tags, selector only for URL, timeout default. Adds value.
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?
Description clearly states the tool runs axe-core accessibility rules on HTML or URL, returns violations grouped by impact with fix guidance and helpUrl links. This is specific and distinguishes it from sibling tools like screenshot or bundler checks.
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?
Description explains two usage modes (HTML string vs live URL) and mentions prerequisite (playwright for URL). Does not explicitly state when not to use or alternatives, but context with siblings makes it clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Walk a build directory, measure raw + gzip (+ optional brotli) sizes per file, and flag files that exceed a global or per-entry KB budget. CI-friendly JSON output.
| Name | Required | Description | Default |
|---|---|---|---|
| buildDir | Yes | Absolute path to the build output directory (e.g. dist, .next/static). | |
| budgetKb | No | Global size budget in KB. Flags any file above this (gzipped). Default: 250. | |
| perEntryBudgetKb | No | Per-file budgets in KB keyed by glob-ish substring match. Matching file is checked against the most specific (longest) matching key. | |
| ext | No | File extensions to include (default: ['.js','.mjs','.cjs','.css']). | |
| includeBrotli | No | Also compute brotli sizes (slower). |
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 explicitly states it walks a directory, measures sizes, and flags exceedances. It also mentions CI-friendly JSON output. While it doesn't disclose potential actions (e.g., read-only or any side effects), the described behavior aligns with a read-only check, making it fairly transparent.
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?
Two sentences, each purposeful. Front-loaded with the core action ('Walk a build directory...'). No wasted words. The structure is optimal for quick agent comprehension.
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?
No output schema is present, so the description should clarify the output format. It mentions 'CI-friendly JSON output' but does not specify structure, exit codes, or how budgets are compared. With 5 parameters and a nested object, more detail would help the agent use it effectively. Score 3 reflects adequate but not complete guidance.
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 coverage is 100%, so baseline is 3. The description adds minimal value beyond the schema: it mentions 'global or per-entry KB budget' but does not explain the glob matching or default values. The schema already describes each parameter adequately. Thus, the description does not significantly enhance parameter understanding.
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 uses a specific verb ('Walk') and resource ('build directory') and clearly states the action: measure raw/gzip/brotli sizes and flag budget violations. This distinguishes it from sibling tools like 'axe_audit' (accessibility) and 'scaffold_react_component', which serve entirely different purposes.
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 clearly implies the tool is for bundle size checks, providing context for use. It does not explicitly state when not to use it or compare with alternatives, but siblings are in unrelated domains, so the context is sufficient. A score of 4 reflects clear context without formal exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Diff two design-token JSON files (W3C DTCG / Style Dictionary style). Reports added, removed, and changed tokens with $type-aware classification. Useful for PR review of design-system changes.
| Name | Required | Description | Default |
|---|---|---|---|
| beforePath | Yes | Path to the baseline tokens JSON (Style Dictionary / W3C DTCG format). | |
| afterPath | Yes | Path to the new tokens JSON. | |
| ignoreKeys | No | Dot-path keys to ignore (e.g. ['$description','$extensions']). |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden of behavioral disclosure. It states the tool reports added, removed, and changed tokens with $type-aware classification, which implies a read-only comparison. It does not mention side effects or required permissions, but for a diff tool, the behavior is sufficiently transparent.
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 only two sentences long, front-loading the core action in the first sentence and adding a use case in the second. Every word is informative with no 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 has no output schema, the description adequately describes what the diff reports (added, removed, changed tokens). It mentions the input format and the classification approach. The absence of output format details is a minor gap, but overall the description is sufficient for an agent to understand the tool's behavior.
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 coverage is 100% with clear descriptions for all three parameters. The tool description adds high-level context (e.g., format, output behavior) but does not provide additional meaning about the parameters beyond the schema. Baseline of 3 is appropriate.
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 diffs design-token JSON files, specifies the format (W3C DTCG / Style Dictionary), and lists what it reports (added, removed, changed tokens with type-aware classification). The use case (PR review) distinguishes it from sibling tools (e.g., axe_audit, bundle_budget_check).
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 explicitly mentions a use case ('Useful for PR review of design-system changes'), which guides when to use it. It does not specify when not to use it or name alternatives, but the context is clear and no sibling tool overlaps with this diffing functionality.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Launch headless Chromium, navigate to a URL, and save a PNG screenshot. Supports element selectors, full-page capture, dark mode, and custom viewports. Requires the optional playwright peer dependency.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | URL to screenshot. | |
| outPath | Yes | Absolute file path to write PNG to. Parent dirs will be created. | |
| selector | No | CSS selector to screenshot instead of the viewport. | |
| fullPage | No | Capture the full scrollable page (default: false). | |
| width | No | Viewport width (default 1280). | |
| height | No | Viewport height (default 800). | |
| deviceScaleFactor | No | DPR (default 2). | |
| colorScheme | No | ||
| waitForSelector | No | Wait for this selector before capturing. | |
| timeoutMs | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must carry behavioral disclosure. It discloses headless Chromium usage, supports for selectors, full-page, dark mode, and viewports. It also mentions the optional Playwright dependency. Missing details on error handling and default behaviors, but still provides significant transparency.
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?
Two sentences, both substantive. First sentence defines core action; second lists key features and a dependency. No redundant or extraneous 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 complexity (10 parameters, no output schema), the description provides a good high-level overview. It specifies the browser engine, supported features, and a dependency. It could mention return behavior (writes file to disk) or default values, but the schema covers defaults. Overall complete given the context.
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 80%, so baseline is 3. The description provides a high-level grouping of features (e.g., 'Supports element selectors, full-page capture, dark mode, and custom viewports') but does not add detail beyond the schema. For the two parameters without schema descriptions (timeoutMs, waitForSelector? waitForSelector has description, timeoutMs does not), the description does not compensate.
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?
Description clearly states 'Launch headless Chromium, navigate to a URL, and save a PNG screenshot.' It uses specific verbs and a concrete resource. Distinct from siblings like axe_audit or bundle_budget_check, which address different tasks.
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 when to use (when a screenshot of a web page is needed) but does not explicitly exclude alternatives or provide when-not-to-use guidance. Siblings are clearly different, so context is clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Write a TypeScript React component, optional test, and optional Storybook story into outDir. Supports functional, forwardRef, and polymorphic (as prop) variants.
| Name | Required | Description | Default |
|---|---|---|---|
| name | Yes | ||
| outDir | Yes | Absolute directory to write the component files into. Created if missing. | |
| variant | No | ||
| withTests | No | ||
| withStory | No | ||
| props | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description must disclose behavioral traits. It mentions file creation and that outDir is created if missing, but does not state whether files are overwritten, side effects, or required permissions. This is insufficient for a write operation.
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 sentence, concise and front-loaded, but could benefit from a second sentence to improve clarity without significant length increase.
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?
With 6 parameters, no output schema, and no annotations, the description is incomplete. It does not explain return value, error conditions, or full behavior of the props parameter.
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 low (17%), and the description does not explain parameters like variant, props, withTests, withStory. It mentions variants but does not elaborate on differences between functional, forwardRef, polymorphic.
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 it writes a TypeScript React component with optional test and Storybook story, and supports three variants. This distinguishes it from sibling tools which are unrelated (auditing, budget, screenshots).
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 basic context but lacks explicit when-to-use or when-not-to-use guidance. Since siblings are in different domains, the need for alternatives is less, but no prerequisites or exclusions are mentioned.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Load a single Storybook story in headless Chromium via iframe.html, optionally screenshot it, and run an axe-core audit against the rendered output. Requires playwright.
| Name | Required | Description | Default |
|---|---|---|---|
| storybookUrl | Yes | Base URL of a running Storybook, e.g. http://localhost:6006. | |
| storyId | Yes | Story id as it appears in the URL (e.g. 'components-button--primary'). | |
| screenshotPath | No | If set, save a PNG of the rendered story to this absolute path. | |
| runAxe | No | Run an axe-core audit of the rendered story (default true). | |
| viewport | No | ||
| colorScheme | No | ||
| timeoutMs | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must cover behavioral traits. It discloses the main actions (load, screenshot, audit) and the need for Playwright, but lacks details on side effects like browser resource usage, error handling, or that it returns axe results.
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?
Two sentences with no fluff. The purpose and key requirement are front-loaded in the first sentence, making it efficient.
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?
The tool has 7 parameters, nested objects, and no output schema. The description omits return values, parameter interactions, and behavioral details like error scenarios or required environment setup, making it incomplete for a complex tool.
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 57%, so the description should add meaning, but it does not explain parameters beyond the schema. It only gives a generic summary, leaving parameter nuances (e.g., viewport, colorScheme, timeoutMs) solely to 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 it loads a single Storybook story via headless Chromium, with optional screenshot and axe-core audit. It distinguishes from siblings like page_screenshot (general) and axe_audit (general audit) by focusing on Storybook stories.
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 for testing Storybook stories but does not explicitly contrast with sibling tools or specify when not to use it. It only mentions a prerequisite (Playwright), not exclusion criteria.
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.
v2.0.0axe_auditbundle_budget_checkdesign_token_diffpage_screenshotscaffold_react_componentstorybook_story_runScored across 6 tools
Each tool targets a distinct frontend concern: accessibility auditing, bundle sizing, design token diffs, page screenshots, component scaffolding, and Storybook testing. No two tools could be confused.
All tool names follow a consistent 'object_action' or 'domain_action' pattern with lowercase underscores (e.g., axe_audit, design_token_diff). No mixing of conventions.
Six tools is an appropriate size for a frontend utility toolkit. Each tool provides a distinct, useful function without being overwhelming or too sparse.
The set covers key frontend tasks: accessibility, performance, design tokens, component generation, and testing. A minor gap is the lack of a linting or formatting tool, but the surface is coherent for its scope.
Live React design-system APIs, patterns, and code validation so AI agents build real UI, not slop.
Code intelligence platform for AI agents. 20 tools for architecture, security & impact analysis.
Direct access to Cypress tests results and accessibility reports in your AI workflow.
Serves your design system and coding standards to coding agents, so they stop guessing.
