@kadoa/mcp 0.7.0-rc.2 → 0.7.0

package/dist/index.js

@@ -17,7 +17,7 @@ var package_default;
 var init_package = __esm(() => {
   package_default = {
     name: "@kadoa/mcp",
-    version: "0.7.0-rc.2",
+    version: "0.7.0",
     description: "Kadoa MCP Server — manage workflows from Claude Desktop, Cursor, and other MCP clients",
     type: "module",
     main: "dist/index.js",
@@ -104,6 +104,11 @@ function createKadoaClient(auth) {
 var MCP_SCRAPE_REQUEST_TIMEOUT_MS = 310000;
 var init_client = () => {};
 
+// src/default-product.ts
+function parseDefaultProductType(value) {
+  return value === "realtime" || value === "workflow" ? value : undefined;
+}
+
 // src/error-summary.ts
 function redactCredentialText(value) {
   return value.replace(/Bearer\s+[^\s,;]+/gi, "Bearer [redacted]").replace(/\b(?:kcs_|tk-)[A-Za-z0-9._~-]+\b/g, "[redacted]").replace(/\beyJ[A-Za-z0-9_-]+\.[A-Za-z0-9_-]+\.[A-Za-z0-9_-]+\b/g, "[redacted]");
@@ -1888,11 +1893,11 @@ function scrapeResult(response, format) {
   };
   return withExtraContent(textOnlyResult(meta), { type: "text", text });
 }
-function registerScrapeTools(server, ctx, helpers) {
+function registerScrapeTools(server, ctx, helpers, defaultProductType) {
   const { withErrorHandling } = helpers;
   server.registerTool("scrape", {
     title: "Scrape URL",
-    description: "Immediately fetch one URL as LLM-friendly markdown or raw HTML without creating a workflow. " + "Markdown is the default. Routing is fully automatic: it starts cheap and escalates through stealth tiers on its own when a site blocks bots, and remembers the working method per host. " + "Use create_workflow instead for structured extraction, recurring runs, monitoring, or navigation-heavy jobs. " + "The result is a JSON metadata block followed by the page text as its own block, capped at 70 KB.",
+    description: "Immediately fetch one URL as LLM-friendly markdown or raw HTML without creating a workflow. " + "Markdown is the default. Routing is fully automatic: it starts cheap and escalates through stealth tiers on its own when a site blocks bots, and remembers the working method per host. " + (defaultProductType === "realtime" ? "Use create_realtime_monitor for monitoring, and create_workflow for one-time or scheduled structured extraction or navigation-heavy jobs. " : "Use create_workflow instead for structured extraction, recurring runs, monitoring, or navigation-heavy jobs. ") + "The result is a JSON metadata block followed by the page text as its own block, capped at 70 KB.",
     inputSchema: strictSchema({
       url: z7.url().refine((value) => {
         const protocol = new URL(value).protocol;
@@ -2161,24 +2166,31 @@ function registerWhoamiTool(server, ctx, helpers) {
   const { withErrorHandling } = helpers;
   server.registerTool("whoami", {
     title: "Show Current User",
-    description: "Show current user details: email, authentication method, and team memberships",
+    description: "Show current user details: email, authentication method, team memberships, and the team's default product",
     inputSchema: strictSchema({}),
     annotations: { readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: false }
   }, withErrorHandling("whoami", async () => {
+    const defaultProductTypePromise = ctx.client.user.getFeatures().then((features) => parseDefaultProductType(features.defaultProductType), () => {
+      return;
+    });
     if (ctx.principal.connector) {
-      const teams = await ctx.client.listTeams();
+      const [teams, defaultProductType] = await Promise.all([
+        ctx.client.listTeams(),
+        defaultProductTypePromise
+      ]);
       const activeTeam = teams.find((team) => team.id === ctx.principal.teamId);
       if (!activeTeam) {
         throw new Error("The active team is unavailable for this credential. Reconnect and choose a team.");
       }
       return jsonResult({
         authMethod: "Shared organization service account",
-        team: { name: activeTeam.name }
+        team: { name: activeTeam.name, ...defaultProductType ? { defaultProductType } : {} }
       });
     }
-    const [user, teams] = await Promise.all([
+    const [user, teams, defaultProductType] = await Promise.all([
       ctx.client.user.getCurrentUser(),
-      ctx.client.listTeams()
+      ctx.client.listTeams(),
+      defaultProductTypePromise
     ]);
     const activeTeam = teams.find((team) => team.id === ctx.principal.teamId);
     if (!activeTeam) {
@@ -2189,7 +2201,8 @@ function registerWhoamiTool(server, ctx, helpers) {
       authMethod: ctx.principal.actor ? `Kadoa Assistant (delegated via ${ctx.principal.actor.sub})` : "Kadoa access token",
       team: {
         name: activeTeam.name,
-        ...activeTeam.memberRole ? { memberRole: activeTeam.memberRole } : {}
+        ...activeTeam.memberRole ? { memberRole: activeTeam.memberRole } : {},
+        ...defaultProductType ? { defaultProductType } : {}
       },
       scopes: ctx.principal.scopes
     });
@@ -2789,7 +2802,7 @@ function summarizeWorkflowAuditEntry(entry) {
 function historyEmptyHint(includeSnapshots) {
   return includeSnapshots ? "A single entry is too large to return inline. Retry without includeSnapshots for the diff summary." : "A single entry is too large to return inline even without snapshots. Retry with a narrower page (e.g. limit=1) or fewer fields.";
 }
-function registerWorkflowCoreTools(server, ctx, helpers) {
+function registerWorkflowCoreTools(server, ctx, helpers, defaultProductType) {
   const { withErrorHandling } = helpers;
   function resolveUrls(args) {
     const urls = args.urls ?? (args.url ? [args.url] : []);
@@ -2897,7 +2910,7 @@ function registerWorkflowCoreTools(server, ctx, helpers) {
 
 ` + "Create a data extraction workflow using agentic navigation. Supports one-time or scheduled runs. " + "If entity and schema are provided, they guide the extraction; otherwise the AI agent auto-detects the schema from the page. " + "The workflow runs asynchronously and may take several minutes. Do NOT poll or sleep-wait for completion. " + `Return the workflow ID to the user and let them check back later with get_workflow or fetch_data.
 
-` + "PREFER TEMPLATES: If the user's request matches an existing template, instantiate it via `templateId` instead of writing a fresh prompt/schema. " + "Use `list_templates` to discover available templates and `get_template` to inspect schemas before deciding. " + "When creating from a template, call this tool with the template's `templateId`, the source `urls`, optional `templateVersion`, and optional `prompt` containing only workflow-specific additional instructions. Never copy the template prompt, entity, or schema into the call, and never silently match or rewrite inline configuration.\n\n" + "NOTE: This tool is for one-time or scheduled extraction ONLY. " + "For continuous realtime monitoring that watches a page for data changes and sends alerts, use the create_realtime_monitor tool instead.",
+` + "PREFER TEMPLATES: If the user's request matches an existing template, instantiate it via `templateId` instead of writing a fresh prompt/schema. " + "Use `list_templates` to discover available templates and `get_template` to inspect schemas before deciding. " + "When creating from a template, call this tool with the template's `templateId`, the source `urls`, optional `templateVersion`, and optional `prompt` containing only workflow-specific additional instructions. Never copy the template prompt, entity, or schema into the call, and never silently match or rewrite inline configuration.\n\n" + "NOTE: This tool is for one-time or scheduled extraction ONLY. " + (defaultProductType === "realtime" ? `This team's default product is realtime monitoring. Do not call this tool unless the user explicitly asked for a one-time run (e.g. "once", "just the current list") or a schedule (e.g. "daily"). A request to get, extract, or scrape data that names no frequency is not explicit: call create_realtime_monitor instead.` : "For continuous realtime monitoring that watches a page for data changes and sends alerts, use the create_realtime_monitor tool instead."),
     inputSchema: strictSchema({
       prompt: extractionPromptInput.describe('Natural language description of what to extract (e.g., "Extract product prices and names"). Required unless templateId is provided. With templateId, use it only for workflow-specific instructions to add to the template prompt.'),
       ...extractionInputShape,
@@ -3050,7 +3063,7 @@ function registerWorkflowCoreTools(server, ctx, helpers) {
   }
   server.registerTool("create_realtime_monitor", {
     title: "Create Realtime Monitor",
-    description: "Create a realtime monitoring workflow that continuously watches a page for data changes and sends alerts. " + "Use this when the user wants to monitor, track, or watch a page in realtime. " + "At least one notification channel (email, webhook, slack, or websocket) is required; notification settings are persisted before Assistant creation is dispatched. " + `Realtime monitors CANNOT be converted to/from one-time or scheduled extraction workflows after creation.
+    description: "Create a realtime monitoring workflow that continuously watches a page for data changes and sends alerts. " + (defaultProductType === "realtime" ? "This team's default product is realtime monitoring: use this tool whenever the user asks to monitor, track, or watch a page, or to get or extract data from it, unless they explicitly ask for a one-time run or a schedule. A request that names no frequency counts as realtime. " : "Use this when the user wants to monitor, track, or watch a page in realtime. ") + "At least one notification channel (email, webhook, slack, or websocket) is required; notification settings are persisted before Assistant creation is dispatched. " + `Realtime monitors CANNOT be converted to/from one-time or scheduled extraction workflows after creation.
 
 ` + "If entity and schema are provided, they guide what the Assistant should monitor; otherwise the Assistant determines what to watch. " + "Creation is asynchronous and the response includes workflow, session, thread, and job IDs plus a dashboard URL. " + "Use generic workflow Assistant tools (get_workflow_assistant, answer_workflow_assistant_question, interrupt_workflow_assistant, resume_workflow_assistant, stop_workflow_assistant) for follow-up status, questions, and controls. " + "Do NOT poll or sleep-wait for completion.",
     inputSchema: strictSchema({
@@ -3727,9 +3740,9 @@ var init_workflows = __esm(() => {
 function registerTools(server, ctx, capabilities) {
   const helpers = { withErrorHandling: createWithErrorHandling(ctx) };
   if (capabilities.scrape)
-    registerScrapeTools(server, ctx, helpers);
+    registerScrapeTools(server, ctx, helpers, capabilities.defaultProductType);
   registerWhoamiTool(server, ctx, helpers);
-  registerWorkflowCoreTools(server, ctx, helpers);
+  registerWorkflowCoreTools(server, ctx, helpers, capabilities.defaultProductType);
   registerAssistantTools(server, ctx, helpers);
   registerWorkflowHistoryTools(server, ctx, helpers);
   registerObservabilityTools(server, ctx, helpers);
@@ -4413,12 +4426,41 @@ var init_request_log = __esm(() => {
 
 // src/server.ts
 import { McpServer as McpServer2 } from "@modelcontextprotocol/server";
+function lifecycleInstruction2(defaultProductType) {
+  return defaultProductType === "realtime" ? "For one-time or scheduled workflows, the lifecycle is create_workflow → get_workflow (check status) → fetch_data (get results). Workflows run asynchronously - never poll or sleep-wait." : "Workflow lifecycle: create_workflow → get_workflow (check status) → fetch_data (get results). Workflows run asynchronously - never poll or sleep-wait.";
+}
+function routingInstructions2(defaultProductType) {
+  if (defaultProductType === "realtime") {
+    return [
+      `This team's default product is realtime monitoring. Use create_realtime_monitor whenever the user asks to monitor, track, or watch a page, or to get or extract data from it, unless they explicitly ask for a one-time run (e.g. "once") or a schedule (e.g. "daily"). A request that names no frequency is not explicit: default to realtime without asking the user to choose. It first persists notification channels, then asynchronously returns workflow/session/thread/job IDs; use the workflow Assistant tools for follow-up status, questions, and controls.`,
+      "Use create_workflow only when the user explicitly asks for a one-time run or a schedule. Use scrape for an immediate raw HTML or markdown fetch from one URL."
+    ];
+  }
+  return [
+    "Use create_realtime_monitor only when the user wants continuous realtime change detection with alerts. It first persists notification channels, then asynchronously returns workflow/session/thread/job IDs; use the workflow Assistant tools for follow-up status, questions, and controls.",
+    "For one-time or scheduled extraction, use create_workflow. Use scrape for an immediate raw HTML or markdown fetch from one URL. Use create_workflow for structured extraction, recurring runs, monitoring, or navigation-heavy jobs."
+  ];
+}
 async function createServer3(auth, options) {
   const ctx = {
     client: createKadoaClient(auth),
     principal: auth.principal,
     teamId: auth.principal.teamId
   };
+  let capabilities = options?.capabilities;
+  if (!capabilities) {
+    try {
+      const response = await ctx.client.user.getFeatures();
+      capabilities = {
+        scrape: response.features.scrape,
+        defaultProductType: parseDefaultProductType(response.defaultProductType)
+      };
+    } catch (error) {
+      const message = error instanceof Error ? error.message : "unknown error";
+      console.error(`[CAPABILITIES] Failed to resolve team capabilities; gated tools will be hidden: ${message}`);
+      capabilities = { scrape: false };
+    }
+  }
   const server = new McpServer2({ name: "kadoa", version: package_default.version }, {
     instructions: [
       "IMPORTANT: One workflow = one source. A single workflow can extract many different fields, tables, and sections from the same URL(s) using a rich schema.",
@@ -4427,10 +4469,9 @@ async function createServer3(auth, options) {
       "Kadoa workflows use agentic navigation: the AI agent can browse pages, click buttons, fill forms, select dropdowns, paginate, open detail pages, and handle multi-step interactions.",
       "Describe the full navigation steps in the prompt (e.g., 'select year 2022 from the dropdown, click Search, extract the table, then repeat for 2023').",
       "",
-      "Workflow lifecycle: create_workflow → get_workflow (check status) → fetch_data (get results). Workflows run asynchronously - never poll or sleep-wait.",
+      lifecycleInstruction2(capabilities.defaultProductType),
       "",
-      "Use create_realtime_monitor only when the user wants continuous realtime change detection with alerts. It first persists notification channels, then asynchronously returns workflow/session/thread/job IDs; use the workflow Assistant tools for follow-up status, questions, and controls.",
-      "For one-time or scheduled extraction, use create_workflow. Use scrape for an immediate raw HTML or markdown fetch from one URL. Use create_workflow for structured extraction, recurring runs, monitoring, or navigation-heavy jobs.",
+      ...routingInstructions2(capabilities.defaultProductType),
       "If a one-time or scheduled workflow fails, retry it with run_workflow using the existing workflow ID. Do NOT delete and recreate the workflow just to retry it - preserve its workflow ID and configuration. Realtime workflows cannot be manually run.",
       "Use list_changes and get_change to retrieve detected diffs from realtime monitoring workflows.",
       "List tools, plus fetch_data, get_workflow_history, and get_workflow_assistant_timeline, all return pagination {totalCount?, hasMore, next?}. When hasMore is true, call the same tool again with pagination.next as its arguments to get the next page. totalCount is present only when the total is known; never treat the page size as the total.",
@@ -4444,17 +4485,6 @@ async function createServer3(auth, options) {
     ].join(`
 `)
   });
-  let capabilities = options?.capabilities;
-  if (!capabilities) {
-    try {
-      const response = await ctx.client.user.getFeatures();
-      capabilities = { scrape: response.features.scrape };
-    } catch (error) {
-      const message = error instanceof Error ? error.message : "unknown error";
-      console.error(`[CAPABILITIES] Failed to resolve team capabilities; gated tools will be hidden: ${message}`);
-      capabilities = { scrape: false };
-    }
-  }
   registerTools(server, ctx, capabilities);
   server.server.onerror = (error) => console.error("[MCP Error]", safeErrorSummary(error));
   return server;
@@ -4534,7 +4564,7 @@ async function requestNeedsCapabilityLookup(request) {
       return false;
     }
   }
-  return method === "tools/list" || method === "tools/call" && name === "scrape";
+  return method === "initialize" || method === "server/discover" || method === "tools/list" || method === "tools/call" && name === "scrape";
 }
 function isClientCausedHandlerError(error) {
   if (error instanceof ProtocolError)
@@ -4874,12 +4904,41 @@ init_package();
 init_client();
 init_tools();
 import { McpServer } from "@modelcontextprotocol/server";
+function lifecycleInstruction(defaultProductType) {
+  return defaultProductType === "realtime" ? "For one-time or scheduled workflows, the lifecycle is create_workflow → get_workflow (check status) → fetch_data (get results). Workflows run asynchronously - never poll or sleep-wait." : "Workflow lifecycle: create_workflow → get_workflow (check status) → fetch_data (get results). Workflows run asynchronously - never poll or sleep-wait.";
+}
+function routingInstructions(defaultProductType) {
+  if (defaultProductType === "realtime") {
+    return [
+      `This team's default product is realtime monitoring. Use create_realtime_monitor whenever the user asks to monitor, track, or watch a page, or to get or extract data from it, unless they explicitly ask for a one-time run (e.g. "once") or a schedule (e.g. "daily"). A request that names no frequency is not explicit: default to realtime without asking the user to choose. It first persists notification channels, then asynchronously returns workflow/session/thread/job IDs; use the workflow Assistant tools for follow-up status, questions, and controls.`,
+      "Use create_workflow only when the user explicitly asks for a one-time run or a schedule. Use scrape for an immediate raw HTML or markdown fetch from one URL."
+    ];
+  }
+  return [
+    "Use create_realtime_monitor only when the user wants continuous realtime change detection with alerts. It first persists notification channels, then asynchronously returns workflow/session/thread/job IDs; use the workflow Assistant tools for follow-up status, questions, and controls.",
+    "For one-time or scheduled extraction, use create_workflow. Use scrape for an immediate raw HTML or markdown fetch from one URL. Use create_workflow for structured extraction, recurring runs, monitoring, or navigation-heavy jobs."
+  ];
+}
 async function createServer(auth, options) {
   const ctx = {
     client: createKadoaClient(auth),
     principal: auth.principal,
     teamId: auth.principal.teamId
   };
+  let capabilities = options?.capabilities;
+  if (!capabilities) {
+    try {
+      const response = await ctx.client.user.getFeatures();
+      capabilities = {
+        scrape: response.features.scrape,
+        defaultProductType: parseDefaultProductType(response.defaultProductType)
+      };
+    } catch (error) {
+      const message = error instanceof Error ? error.message : "unknown error";
+      console.error(`[CAPABILITIES] Failed to resolve team capabilities; gated tools will be hidden: ${message}`);
+      capabilities = { scrape: false };
+    }
+  }
   const server = new McpServer({ name: "kadoa", version: package_default.version }, {
     instructions: [
       "IMPORTANT: One workflow = one source. A single workflow can extract many different fields, tables, and sections from the same URL(s) using a rich schema.",
@@ -4888,10 +4947,9 @@ async function createServer(auth, options) {
       "Kadoa workflows use agentic navigation: the AI agent can browse pages, click buttons, fill forms, select dropdowns, paginate, open detail pages, and handle multi-step interactions.",
       "Describe the full navigation steps in the prompt (e.g., 'select year 2022 from the dropdown, click Search, extract the table, then repeat for 2023').",
       "",
-      "Workflow lifecycle: create_workflow → get_workflow (check status) → fetch_data (get results). Workflows run asynchronously - never poll or sleep-wait.",
+      lifecycleInstruction(capabilities.defaultProductType),
       "",
-      "Use create_realtime_monitor only when the user wants continuous realtime change detection with alerts. It first persists notification channels, then asynchronously returns workflow/session/thread/job IDs; use the workflow Assistant tools for follow-up status, questions, and controls.",
-      "For one-time or scheduled extraction, use create_workflow. Use scrape for an immediate raw HTML or markdown fetch from one URL. Use create_workflow for structured extraction, recurring runs, monitoring, or navigation-heavy jobs.",
+      ...routingInstructions(capabilities.defaultProductType),
       "If a one-time or scheduled workflow fails, retry it with run_workflow using the existing workflow ID. Do NOT delete and recreate the workflow just to retry it - preserve its workflow ID and configuration. Realtime workflows cannot be manually run.",
       "Use list_changes and get_change to retrieve detected diffs from realtime monitoring workflows.",
       "List tools, plus fetch_data, get_workflow_history, and get_workflow_assistant_timeline, all return pagination {totalCount?, hasMore, next?}. When hasMore is true, call the same tool again with pagination.next as its arguments to get the next page. totalCount is present only when the total is known; never treat the page size as the total.",
@@ -4905,17 +4963,6 @@ async function createServer(auth, options) {
     ].join(`
 `)
   });
-  let capabilities = options?.capabilities;
-  if (!capabilities) {
-    try {
-      const response = await ctx.client.user.getFeatures();
-      capabilities = { scrape: response.features.scrape };
-    } catch (error) {
-      const message = error instanceof Error ? error.message : "unknown error";
-      console.error(`[CAPABILITIES] Failed to resolve team capabilities; gated tools will be hidden: ${message}`);
-      capabilities = { scrape: false };
-    }
-  }
   registerTools(server, ctx, capabilities);
   server.server.onerror = (error) => console.error("[MCP Error]", safeErrorSummary(error));
   return server;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@kadoa/mcp",
-  "version": "0.7.0-rc.2",
+  "version": "0.7.0",
   "description": "Kadoa MCP Server — manage workflows from Claude Desktop, Cursor, and other MCP clients",
   "type": "module",
   "main": "dist/index.js",