@synergy-design-system/mcp 3.6.0 → 3.8.0

package/CHANGELOG.md

@@ -1,5 +1,35 @@
 # Changelog
 
+## 3.8.0
+
+### Minor Changes
+
+- [#1276](https://github.com/synergy-design-system/synergy-design-system/pull/1276) [`a6b68f2`](https://github.com/synergy-design-system/synergy-design-system/commit/a6b68f2126ee0c63f2e9f5a91d96e97c2486e3c9) Thanks [@schilchSICKAG](https://github.com/schilchSICKAG)! - Released on: 2026-05-21
+
+  feat: ✨ component context guidelines ([#1187](https://github.com/synergy-design-system/synergy-design-system/issues/1187))
+  - Correct `@since` flags for all components
+  - Updated component usage rules (beta)
+  - Add missing rules files
+  - Create overviews for all components
+
+### Patch Changes
+
+- Updated dependencies [[`a6b68f2`](https://github.com/synergy-design-system/synergy-design-system/commit/a6b68f2126ee0c63f2e9f5a91d96e97c2486e3c9)]:
+  - @synergy-design-system/metadata@3.8.0
+
+## 3.7.0
+
+### Minor Changes
+
+- Released on: 2026-05-11
+
+  chore: ✨ Update Metadata and MCP with latest metadata
+
+### Patch Changes
+
+- Updated dependencies [[`9cf37bc`](https://github.com/synergy-design-system/synergy-design-system/commit/9cf37bcd8c2579946e5c314a246a80a443ec8357)]:
+  - @synergy-design-system/metadata@3.7.0
+
 ## 3.6.0
 
 ### Minor Changes

package/README.md

@@ -145,6 +145,12 @@ Example:
   // "toon": Encode structured data to compact toon text format
   "compression": "none",
 
+  // Experimental feature toggles
+  "experimentalFeatures": {
+    // Enables experimental intent policy tools
+    "intentTools": true,
+  },
+
   // Default parameters for each endpoint can be overridden
   "tools": {
     "assetInfo": {
@@ -156,11 +162,36 @@ Example:
       // Defines which type of information to return.
       // full = filtered source files,
       // examples = markdown examples,
-      // interface = markdown API overview.
+      // interface = markdown API overview,
+      // rules = markdown usage, design, and accessibility guidance.
       // Note that examples and interface are only available
       // for vanilla components at the moment.
       "layer": "full",
     },
+    "intentCategoriesList": {
+      // Default phases for intent-categories-list
+      "includePhases": ["experimental"],
+    },
+    "intentComponentGuide": {
+      // Framework profile used when omitted in tool call
+      "framework": "vanilla",
+      "includePhases": ["experimental"],
+    },
+    "intentComponentValidate": {
+      "framework": "vanilla",
+      "includePhases": ["experimental"],
+    },
+    "intentOptions": {
+      "framework": "vanilla",
+      "includeDiagnostics": false,
+      "includePhases": ["experimental"],
+      "maxAlternatives": 5,
+    },
+    "intentTaskRecommendations": {
+      "framework": "vanilla",
+      "includePhases": ["experimental"],
+      "maxAlternatives": 5,
+    },
     "tokenInfo": {
       // If you are preferring scss, use "sass" here
       "type": "css",
@@ -169,6 +200,33 @@ Example:
 }
 ```
 
+#### Enabling Experimental Intent Endpoints
+
+Intent tools are experimental and disabled by default.
+Enable them explicitly with:
+
+```jsonc
+{
+  "experimentalFeatures": {
+    "intentTools": true,
+  },
+}
+```
+
+When enabled, these additional features become available:
+
+#### Resources
+
+- `synergy://intent-categories/list`
+
+#### Tools
+
+- `intent-categories-list`
+- `intent-component-guide`
+- `intent-component-validate`
+- `intent-task-recommendations`
+- `intent-options`
+
 #### CLI Override Precedence
 
 CLI flags take precedence over configuration file values:
@@ -301,7 +359,9 @@ This lets you change per-tool defaults without modifying the MCP server code.
 
 ## Available Resources
 
-The MCP server currently registers 5 resources. Resources expose static, read-only data that does not change during server runtime. Clients that support MCP resources can read them directly without calling a tool.
+The MCP server currently registers 5 stable resources by default. Resources expose static, read-only data that does not change during server runtime. Clients that support MCP resources can read them directly without calling a tool.
+
+When `experimentalFeatures.intentTools` is enabled, 1 additional intent resource is registered.
 
 Resource identifier reference (exact values used by the server):
 
@@ -310,6 +370,7 @@ Resource identifier reference (exact values used by the server):
 - `synergy://component-clusters/list` → name: `component-clusters-list`
 - `synergy://styles/list` → name: `styles-list`
 - `synergy://templates/list` → name: `templates-list`
+- `synergy://intent-categories/list` → name: `intent-categories-list` (experimental)
 
 ### 1. `synergy://components/list`
 
@@ -397,9 +458,33 @@ Resource identifier reference (exact values used by the server):
 ["app-shell", "dashboard", "form", ...]
 ```
 
+### 6. `synergy://intent-categories/list` (experimental)
+
+**Name:** `intent-categories-list`
+
+**MIME type:** `application/json`
+
+**Description:** Available intent categories in the Synergy intent policy layer.
+
+**Registration:** Only available when `experimentalFeatures.intentTools` is `true`.
+
+**Example:**
+
+```json
+{
+  "data": [
+    {
+      "description": "User actions and commands",
+      "id": "action",
+      "label": "Action"
+    }
+  ]
+}
+```
+
 ## Available Tools
 
-The MCP server currently registers 16 tools.
+The MCP server currently registers 16 stable tools by default.
 
 ### 1. `component-list`
 
@@ -450,13 +535,14 @@ Example prompts:
 
 - `component` (string, required): The component name. Must start with `syn-`, for example `syn-button`.
 - `framework` (string, optional): `react`, `vue`, `angular`, or `vanilla`. Defaults to the runtime config value, which is `vanilla` by default.
-- `layer` (string, optional): `full`, `examples`, or `interface`. Defaults to the runtime config value, which is `full` by default. `examples` and `interface` are currently only available for vanilla components.
+- `layer` (string, optional): `full`, `examples`, `interface`, or `rules`. Defaults to the runtime config value, which is `full` by default. `examples` and `interface` are currently only available for vanilla components. `rules` returns component usage, design, and accessibility guidance.
 
 **Example prompts:**
 
 - "How do I use syn-button in React?"
 - "Show me the interface docs for syn-dialog"
 - "Give me examples for syn-card"
+- "Show me the rules for syn-accordion"
 
 ### 4. `asset-list`
 
@@ -649,6 +735,110 @@ Example prompts:
 - "Show me the setup instructions for tokens"
 - "Give me the Synergy assets setup and limitations"
 
+### Experimental Tools (Intent Policy)
+
+The following endpoints are experimental and are only registered when `experimentalFeatures.intentTools` is set to `true`.
+
+### 17. `intent-categories-list` (experimental)
+
+**Description:** List available intent categories in the intent policy layer.
+
+**Parameters:**
+
+- `includePhases` (array, optional): Intent phases to include. Defaults to runtime config `tools.intentCategoriesList.includePhases` (`["experimental"]` by default).
+
+**Example prompts:**
+
+- "List intent categories"
+- "Show experimental intent categories"
+
+### 18. `intent-component-guide` (experimental)
+
+**Description:** Answer the question: What can I do with a component in the intent system?
+
+**Parameters:**
+
+- `component` (string, required): Component tag, for example `syn-button`.
+- `framework` (string, optional): `react-wrapper`, `react-web-components`, `angular`, `vue`, or `vanilla`. Defaults to runtime config `tools.intentComponentGuide.framework`.
+- `includePhases` (array, optional): Defaults to runtime config `tools.intentComponentGuide.includePhases`.
+
+**Example prompts:**
+
+- "What can I do with syn-button?"
+- "Show intent guide for syn-button in react-web-components"
+
+### 19. `intent-component-validate` (experimental)
+
+**Description:** Answer the question: Do I use a component correctly for a specific intent?
+
+**Parameters:**
+
+- `component` (string, required): Component tag, for example `syn-button`.
+- `intent` (string, required): Intent id, for example `action.submit`.
+- `markup` (string, required): Template/markup source to lint. The tool derives structure internally.
+- `framework` (string, optional): Defaults to runtime config `tools.intentComponentValidate.framework`.
+- `includePhases` (array, optional): Defaults to runtime config `tools.intentComponentValidate.includePhases`.
+
+**Example prompts:**
+
+- "Do I use syn-button right for action.submit?"
+- "Validate this syn-button markup for action.submit: <syn-button type=\"submit\" variant=\"filled\">Send</syn-button>"
+
+### 20. `intent-task-recommendations` (experimental)
+
+**Description:** Answer the question: What does Synergy provide for a specific task intent?
+
+**Parameters:**
+
+- `taskId` (string, required): Intent id representing the task.
+- `framework` (string, optional): Defaults to runtime config `tools.intentTaskRecommendations.framework`.
+- `includePhases` (array, optional): Defaults to runtime config `tools.intentTaskRecommendations.includePhases`.
+- `maxAlternatives` (number, optional): Defaults to runtime config `tools.intentTaskRecommendations.maxAlternatives`.
+- `preferredTargets` (array, optional): Preferred target ids.
+- `avoidTargets` (array, optional): Target ids to avoid.
+- `content` (string, optional): Optional snippet content.
+
+**Example prompts:**
+
+- "What does Synergy provide to submit a form?"
+- "Recommend components for action.submit"
+
+### 21. `intent-options` (experimental)
+
+**Description:** Answer the question: What are my renderable options for a specific intent?
+
+**Parameters:**
+
+- `intentId` (string, required): Intent id to resolve.
+- `framework` (string, optional): Defaults to runtime config `tools.intentOptions.framework`.
+- `includePhases` (array, optional): Defaults to runtime config `tools.intentOptions.includePhases`.
+- `includeDiagnostics` (boolean, optional): Defaults to runtime config `tools.intentOptions.includeDiagnostics`.
+- `maxAlternatives` (number, optional): Defaults to runtime config `tools.intentOptions.maxAlternatives`.
+- `content` (string, optional): Optional preview content.
+
+**Example prompts:**
+
+- "What are my renderable options for navigation.link-list.grouped?"
+- "Show intent options for action.submit"
+
+## Available Prompts
+
+The MCP server currently registers 1 prompt.
+
+### 1. `explain-component-rules`
+
+**Description:** Explains the usage rules, design guidelines, and accessibility considerations for a Synergy component.
+
+**Parameters:**
+
+- `component` (string, required): The component name. Must start with `syn-`, for example `syn-button`.
+
+**Example prompts:**
+
+- "Explain the rules for syn-button"
+- "What are the design and accessibility guidelines for syn-dialog?"
+- "Give me the usage guidance for syn-accordion"
+
 ## Usage Examples
 
 ### Command Line Interface
@@ -741,6 +931,7 @@ Once connected to an AI assistant, you can use prompts like:
 
 ```text
 Show me how to use syn-button in React
+Explain the rules for syn-accordion
 Give me the interface docs for syn-select
 What token formats are available?
 How do I set up Synergy for Vue?
@@ -759,7 +950,7 @@ src/
 ├── bin/
 │   ├── clean.js         # Removes dist/ before builds
 │   └── start.ts         # CLI entry point for syn-mcp
-├── server.ts            # MCP server creation, tool and resource registration
+├── server.ts            # MCP server creation, tool, prompt and resource registration
 ├── middleware/          # Tool execution middleware pipeline
 │   ├── compose.ts       # composeMiddlewares (reduceRight composition)
 │   ├── compression.ts   # withCompressionMiddleware (experimental)
@@ -767,10 +958,14 @@ src/
 │   ├── logging.ts       # withToolLoggingMiddleware
 │   ├── types.ts         # ToolMiddleware, ToolMiddlewareContext, RawToolHandler, WithErrorHandlerOptions
 │   └── index.ts         # Middleware module entrypoint
+├── prompts/             # MCP prompt implementations
+│   ├── component-rules.ts
+│   └── index.ts
 ├── resources/           # MCP resource implementations (static, read-only data)
 │   ├── component-list.ts
 │   ├── asset-list.ts
 │   ├── component-cluster-list.ts
+│   ├── intent-categories-list.ts
 │   ├── styles-list.ts
 │   ├── templates-list.ts
 │   └── index.ts
@@ -782,6 +977,11 @@ src/
 │   ├── component-list.ts
 │   ├── davinci-migration-info.ts
 │   ├── davinci-migration-list.ts
+│   ├── intent-categories-list.ts
+│   ├── intent-component-guide.ts
+│   ├── intent-component-validate.ts
+│   ├── intent-options.ts
+│   ├── intent-task-recommendations.ts
 │   ├── migration-info.ts
 │   ├── migration-list.ts
 │   ├── setup.ts
@@ -793,20 +993,8 @@ src/
 │   ├── tokens-list.ts
 │   └── index.ts
 ├── transports/          # Transport factory and implementations
-│   ├── http.ts
-│   ├── stdio.ts
-│   └── index.ts
 ├── types/               # Shared type definitions
-│   └── tool-response.ts
-└── utilities/           # Runtime config, metadata adapters, and CLI helpers
-  ├── cli.ts
-  ├── config.ts
-  ├── davinci.ts
-  ├── metadata.ts
-  ├── migration.ts
-  ├── rules.ts
-  ├── server.ts
-  └── index.ts
+└── utilities/           # Runtime config, metadata adapters, intent defaults, and CLI helpers
 rules/                   # Markdown guidance files prepended to selected tool output
 test/
 ├── e2e/                 # End-to-end MCP tests
@@ -913,6 +1101,7 @@ The MCP server is intentionally small:
 - `src/bin/start.ts` parses CLI arguments, loads optional runtime config, resolves overrides, and starts the selected transport.
 - `src/transports/` contains the transport factory and runtime implementations for stdio and HTTP/HTTPS.
 - `src/server.ts` creates the `McpServer` instance and registers all exported tools from `src/tools/index.ts` and all exported resources from `src/resources/index.ts`.
+- `src/server.ts` applies feature-flag gating generically: exports whose factory names start with `intent` are only registered when `experimentalFeatures.intentTools` is enabled.
 - Tool implementations in `src/tools/` call the public APIs of `@synergy-design-system/metadata` to retrieve data.
 - Resource implementations in `src/resources/` expose static, read-only data that does not change during server runtime. Resources bypass the tool middleware pipeline entirely — no compression, logging, or error wrapping is applied.
 - Utilities in `src/utilities/` handle runtime config, MCP response shaping, DaVinci migration extraction, and package migration document loading.

package/dist/middleware/index.d.ts

@@ -7,3 +7,4 @@ export { composeMiddlewares } from './compose.js';
 export { withCompressionMiddleware } from './compression.js';
 export { withErrorHandlingMiddleware } from './error-handler.js';
 export { withToolLoggingMiddleware } from './logging.js';
+export { withPromptLoggingMiddleware } from './logging.js';

package/dist/middleware/index.js

@@ -6,3 +6,4 @@ export { composeMiddlewares } from './compose.js';
 export { withCompressionMiddleware } from './compression.js';
 export { withErrorHandlingMiddleware } from './error-handler.js';
 export { withToolLoggingMiddleware } from './logging.js';
+export { withPromptLoggingMiddleware } from './logging.js';

package/dist/middleware/logging.d.ts

@@ -1,8 +1,3 @@
-/**
- * Logging middleware for tool execution.
- * Records tool call duration, token count, success/error state, and parameters.
- * Respects logging config: early-exits if logging is disabled to avoid token counting overhead.
- */
 import type { ToolMiddleware } from './types.js';
 /**
  * Middleware that logs details about tool execution, including duration, success/failure, token count, and parameters.
@@ -12,3 +7,8 @@ import type { ToolMiddleware } from './types.js';
  * @returns A new handler function that wraps the original tool logic with logging functionality.
  */
 export declare const withToolLoggingMiddleware: ToolMiddleware<Record<string, unknown>>;
+/**
+ * Middleware that logs prompt execution details.
+ * Token counting is based on the resolved prompt text payload.
+ */
+export declare const withPromptLoggingMiddleware: ToolMiddleware<Record<string, unknown>>;

package/dist/middleware/logging.js

@@ -6,6 +6,31 @@
 import { logToolCall } from '../utilities/logger.js';
 import { countTextTokens, toTextPayload, } from '../utilities/token-counter.js';
 import { toContentArray } from '../utilities/metadata.js';
+const isPromptResponse = (value) => {
+    if (!value || typeof value !== 'object') {
+        return false;
+    }
+    const maybePrompt = value;
+    return typeof maybePrompt.description === 'string' && Array.isArray(maybePrompt.messages);
+};
+const toPromptTextPayload = (result) => {
+    const lines = result.flatMap((entry) => {
+        if (typeof entry === 'string') {
+            return [entry];
+        }
+        if (isPromptResponse(entry)) {
+            return entry.messages
+                .map(message => {
+                if (message.content.type === 'text')
+                    return message.content.text;
+                return '';
+            })
+                .filter(Boolean);
+        }
+        return [JSON.stringify(entry)];
+    });
+    return lines.join('\n\n');
+};
 /**
  * Middleware that logs details about tool execution, including duration, success/failure, token count, and parameters.
  * It checks the logging configuration and skips logging (and token counting) if logging is disabled to optimize performance.
@@ -62,3 +87,50 @@ export const withToolLoggingMiddleware = (next, context) => async (args) => {
         });
     }
 };
+/**
+ * Middleware that logs prompt execution details.
+ * Token counting is based on the resolved prompt text payload.
+ */
+export const withPromptLoggingMiddleware = (next, context) => async (args) => {
+    if (context.config.logging.localFile.path === null) {
+        return next(args);
+    }
+    const startedAt = process.hrtime.bigint();
+    let success = false;
+    let errorMessage;
+    let tokenCount;
+    try {
+        const result = await next(args);
+        success = true;
+        const textPayload = toPromptTextPayload(result);
+        tokenCount = await countTextTokens(textPayload);
+        return result;
+    }
+    catch (error) {
+        errorMessage = error instanceof Error ? error.message : String(error);
+        throw error;
+    }
+    finally {
+        const durationMs = Number(process.hrtime.bigint() - startedAt) / 1_000_000;
+        const internalProps = new Set([
+            'requestId',
+            'signal',
+            'requestInfo',
+            'sessionId',
+        ]);
+        const loggableParameters = {};
+        Object.entries(args).forEach(([key, value]) => {
+            if (!internalProps.has(key)) {
+                loggableParameters[key] = value;
+            }
+        });
+        await logToolCall({
+            durationMs,
+            errorMessage,
+            parameters: loggableParameters,
+            success,
+            tokenCount,
+            toolName: `prompt:${context.toolName}`,
+        });
+    }
+};

package/dist/prompts/component-rules.d.ts

@@ -0,0 +1,2 @@
+import type { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
+export declare const explainComponentRules: (server: McpServer) => void;

package/dist/prompts/component-rules.js

@@ -0,0 +1,36 @@
+import { z } from 'zod';
+import { getRulesForComponent, } from '@synergy-design-system/metadata';
+import { promptHandler, } from '../utilities/index.js';
+export const explainComponentRules = (server) => {
+    server.registerPrompt('explain-component-rules', {
+        argsSchema: {
+            component: z
+                .string()
+                .startsWith('syn-')
+                .describe('The name of the component to get rules for.'),
+        },
+        description: 'Explains the usage rules, design guidelines, and accessibility considerations for a Synergy component.',
+        title: 'Explain component rules',
+    }, promptHandler('explain-component-rules', async ({ component }) => {
+        const metadata = await getRulesForComponent(component);
+        const rulesEntries = metadata.data?.rules ?? [];
+        const warningsList = metadata.data?.warnings ?? [];
+        const rulesMarkdown = rulesEntries.map((entry) => entry.content).join('\n\n');
+        const warnings = warningsList.length > 0
+            ? `\n\nWarnings:\n${warningsList.map((warning) => `- ${warning}`).join('\n')}`
+            : '';
+        return [
+            `
+You are an expert on the Synergy Design System.
+
+Base the explanation strictly on the provided data as seen below.
+Do not speculate or invent any rules.
+Structure the information clearly for a human reader.`,
+            rulesMarkdown && rulesMarkdown.trim().length > 0
+                ? `${rulesMarkdown}${warnings}`
+                : `No rules markdown is available for ${component}.${warnings}`,
+        ];
+    }, {
+        description: ({ component }) => `Explain design system rules for ${component}`,
+    }));
+};

package/dist/prompts/index.d.ts

@@ -0,0 +1 @@
+export * from './component-rules.js';

package/dist/prompts/index.js

@@ -0,0 +1 @@
+export * from './component-rules.js';

package/dist/resources/index.d.ts

@@ -1,5 +1,6 @@
 export * from './asset-list.js';
 export * from './component-cluster-list.js';
 export * from './component-list.js';
+export * from './intent-categories-list.js';
 export * from './styles-list.js';
 export * from './templates-list.js';

package/dist/resources/index.js

@@ -1,5 +1,6 @@
 export * from './asset-list.js';
 export * from './component-cluster-list.js';
 export * from './component-list.js';
+export * from './intent-categories-list.js';
 export * from './styles-list.js';
 export * from './templates-list.js';

package/dist/resources/intent-categories-list.d.ts

@@ -0,0 +1,7 @@
+import type { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
+/**
+ * Registers a static MCP resource that lists all available intent categories in the Synergy Design System.
+ * Clients that support resources can read this directly without calling a tool.
+ * @param server - The MCP server instance to register the resource on.
+ */
+export declare const intentCategoriesListResource: (server: McpServer) => void;

package/dist/resources/intent-categories-list.js

@@ -0,0 +1,25 @@
+import { experimental_listIntentCategories as listIntentCategories, } from '@synergy-design-system/metadata';
+const RESOURCE_URI = 'synergy://intent-categories/list';
+/**
+ * Registers a static MCP resource that lists all available intent categories in the Synergy Design System.
+ * Clients that support resources can read this directly without calling a tool.
+ * @param server - The MCP server instance to register the resource on.
+ */
+export const intentCategoriesListResource = (server) => {
+    server.registerResource('intent-categories-list', RESOURCE_URI, {
+        description: 'Available intent categories in the Synergy Design System.',
+        mimeType: 'application/json',
+        title: 'Available intent categories',
+    }, async (_uri) => {
+        const categories = await listIntentCategories();
+        return {
+            contents: [
+                {
+                    mimeType: 'application/json',
+                    text: JSON.stringify(categories, null, 2),
+                    uri: RESOURCE_URI,
+                },
+            ],
+        };
+    });
+};

package/dist/server.js

@@ -1,24 +1,41 @@
 import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
 import * as resources from './resources/index.js';
+import * as prompts from './prompts/index.js';
 import * as tools from './tools/index.js';
 import { getVersion } from './utilities/cli.js';
+import { getRuntimeConfig } from './utilities/config.js';
+/**
+ * Determines if a tool, prompt, or resource should be registered based on feature flags.
+ * Items with names starting with 'intent' are gated by the intentTools experimental flag.
+ */
+const shouldRegisterItem = (name, config) => {
+    // Make sure to only load intent-related items if the intentTools experimental feature is enabled
+    if (name.toLowerCase().startsWith('intent')) {
+        return config.experimentalFeatures?.intentTools === true;
+    }
+    return true;
+};
 /**
  * Creates a new instance of the MCP server configured for the Synergy Design System.
  * @returns A new instance of the MCP server configured for the Synergy Design System.
  */
 export const createServer = () => {
     const version = getVersion();
+    const config = getRuntimeConfig();
     const server = new McpServer({
         description: 'A server for the Synergy Design System that provides tools to interact with components and resources.',
         name: 'synergy design system',
         title: 'Synergy Design System MCP Server',
         version,
     });
-    Object.values(tools).forEach(tool => {
-        tool(server);
-    });
-    Object.values(resources).forEach(resource => {
-        resource(server);
-    });
+    Object.entries(prompts)
+        .filter(([name]) => shouldRegisterItem(name, config))
+        .forEach(([, prompt]) => prompt(server));
+    Object.entries(tools)
+        .filter(([name]) => shouldRegisterItem(name, config))
+        .forEach(([, tool]) => tool(server));
+    Object.entries(resources)
+        .filter(([name]) => shouldRegisterItem(name, config))
+        .forEach(([, resource]) => resource(server));
     return server;
 };

package/dist/tools/component-info.js

@@ -12,9 +12,9 @@ export const componentInfoTool = (server) => {
         inputSchema: {
             component: z.string().startsWith('syn-').describe('The name of the component to get information about.'),
             framework: z.enum(['react', 'vue', 'angular', 'vanilla']).optional().describe('The framework of the component, e.g., "react", "vue", etc.'),
-            layer: z.enum(['full', 'examples', 'interface'])
+            layer: z.enum(['full', 'examples', 'interface', 'rules'])
                 .optional()
-                .describe('Defines which type of information to return. full = filtered source files (useful for debugging), examples = markdown examples, interface = markdown API overview. Examples and interface are only available for vanilla components at the moment.'),
+                .describe('Defines which type of information to return. full = filtered source files (useful for debugging), examples = markdown examples, interface = markdown API overview, rules = Usage guidelines and rules for the component. Examples and interface are only available for vanilla components at the moment.'),
         },
         title: 'Component info',
     }, toolHandler('component-info', async ({ component, framework, layer, }) => {
@@ -39,6 +39,9 @@ export const componentInfoTool = (server) => {
             case 'examples':
                 finalContent = metadata.data.examples?.map((entry) => entry.content) ?? [];
                 break;
+            case 'rules':
+                finalContent = metadata.data.rules?.map((entry) => entry.content) ?? [];
+                break;
             default:
                 finalContent = [metadata.data];
         }

package/dist/tools/index.d.ts

@@ -7,6 +7,11 @@ export * from './davinci-migration-info.js';
 export * from './davinci-migration-list.js';
 export * from './migration-info.js';
 export * from './migration-list.js';
+export * from './intent-categories-list.js';
+export * from './intent-component-guide.js';
+export * from './intent-component-validate.js';
+export * from './intent-task-recommendations.js';
+export * from './intent-options.js';
 export * from './setup.js';
 export * from './styles-info.js';
 export * from './styles-list.js';