mcp-researchpowerpack-http 4.0.0 → 4.1.0

package/dist/mcp-use.json

@@ -1,7 +1,7 @@
 {
   "includeInspector": false,
-  "buildTime": "2026-04-10T05:41:54.660Z",
-  "buildId": "81a3d1a6e86f5bde",
+  "buildTime": "2026-04-10T06:16:46.859Z",
+  "buildId": "ace839805d212563",
   "entryPoint": "dist/index.js",
   "widgets": {}
 }

package/dist/src/schemas/reddit.js

@@ -1,35 +1,32 @@
 import { z } from "zod";
 const searchRedditParamsSchema = z.object({
   queries: z.array(
-    z.string().min(1, { message: "search-reddit: Query cannot be empty" }).describe('A Reddit search query. Do NOT add "site:reddit.com" \u2014 it is appended automatically.')
-  ).min(1, { message: "search-reddit: At least 1 query is required" }).max(50, { message: "search-reddit: Maximum 50 queries allowed" }).describe('Array of 1-50 search queries. Each query gets "site:reddit.com" appended and is sent to Serper as a standard Google search. Returns a flat list of deduplicated Reddit URLs. Use with get-reddit-post to fetch full content and extract insights.')
-}).strict();
-const getRedditPostParamsSchema = z.object({
-  urls: z.array(
-    z.string().url({ message: "get-reddit-post: Each URL must be valid" }).describe('A full Reddit post URL (e.g., "https://www.reddit.com/r/subreddit/comments/id/title/").')
-  ).min(1, { message: "get-reddit-post: At least 1 Reddit post URL is required" }).max(50, { message: "get-reddit-post: Maximum 50 Reddit post URLs allowed" }).describe("Array of 1-50 Reddit post URLs. Each post is fetched with full comment trees, then the LLM extracts insights per what_to_extract. Best used after search-reddit."),
-  fetch_comments: z.boolean().default(true).describe("Fetch threaded comment trees for each post. Defaults to true. Comments include author, score, OP markers, and nested replies. Set false only when you need post titles/selftext without community discussion."),
-  what_to_extract: z.string({ error: "get-reddit-post: what_to_extract is required" }).min(5, { message: "get-reddit-post: what_to_extract must be at least 5 characters" }).max(1e3, { message: "get-reddit-post: what_to_extract is too long (max 1000 characters)" }).describe('REQUIRED. Extraction instructions for the LLM. Describes what insights, opinions, or data to pull from each post and its comments. Use pipe separators for multiple targets: "Extract recommendations | pain points | consensus on best practices | specific tools mentioned".')
+    z.string().min(1, { message: "search-reddit: Query cannot be empty" }).describe('A Reddit search query. "site:reddit.com" is appended automatically.')
+  ).min(1, { message: "search-reddit: At least 1 query required" }).max(100, { message: "search-reddit: Maximum 100 queries allowed" }).describe("Search queries for Reddit. Each query is automatically scoped to reddit.com via Google. Returns deduplicated Reddit post URLs.")
 }).strict();
 const searchRedditOutputSchema = z.object({
-  content: z.string().describe("Newline-separated list of unique Reddit URLs discovered across all queries."),
+  content: z.string().describe("Newline-separated list of unique Reddit URLs."),
   metadata: z.object({
-    query_count: z.number().int().nonnegative().describe("Number of queries executed."),
-    total_urls: z.number().int().nonnegative().describe("Total unique Reddit URLs returned.")
-  }).strict().describe("Metadata about the Reddit URL search.")
+    total_items: z.number().int().nonnegative().describe("Number of queries executed."),
+    successful: z.number().int().nonnegative().describe("Queries that returned results."),
+    failed: z.number().int().nonnegative().describe("Queries that failed."),
+    execution_time_ms: z.number().int().nonnegative().describe("Wall clock time in milliseconds.")
+  }).strict()
+}).strict();
+const getRedditPostParamsSchema = z.object({
+  urls: z.array(
+    z.string().url({ message: "get-reddit-post: Each URL must be valid" }).describe("A Reddit post URL.")
+  ).min(1, { message: "get-reddit-post: At least 1 URL required" }).max(100, { message: "get-reddit-post: Maximum 100 URLs allowed" }).describe("Reddit post URLs to fetch. Each post is returned with its full threaded comment tree.")
 }).strict();
 const getRedditPostOutputSchema = z.object({
-  content: z.string().describe("LLM-synthesized extraction from Reddit posts and comments, structured per what_to_extract instructions."),
+  content: z.string().describe("Raw Reddit posts with threaded comments including author, score, and OP markers."),
   metadata: z.object({
-    total_urls: z.number().int().nonnegative().describe("Total number of Reddit post URLs processed."),
-    successful: z.number().int().nonnegative().describe("Number of posts fetched successfully."),
-    failed: z.number().int().nonnegative().describe("Number of post fetches that failed."),
-    fetch_comments: z.boolean().describe("Whether comments were fetched for each post."),
-    total_words_used: z.number().int().nonnegative().describe("Total words used across all posts."),
-    llm_failures: z.number().int().nonnegative().describe("Count of posts where LLM extraction failed (raw content returned instead)."),
-    total_batches: z.number().int().nonnegative().describe("Number of Reddit API batches executed."),
-    rate_limit_hits: z.number().int().nonnegative().describe("Observed Reddit API rate-limit retries during the batch.")
-  }).strict().describe("Metadata about the Reddit post fetch and extraction.")
+    total_items: z.number().int().nonnegative().describe("Number of URLs processed."),
+    successful: z.number().int().nonnegative().describe("Posts fetched successfully."),
+    failed: z.number().int().nonnegative().describe("Posts that failed to fetch."),
+    execution_time_ms: z.number().int().nonnegative().describe("Wall clock time in milliseconds."),
+    rate_limit_hits: z.number().int().nonnegative().describe("Reddit API rate-limit retries.")
+  }).strict()
 }).strict();
 export {
   getRedditPostOutputSchema,

package/dist/src/schemas/reddit.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../src/schemas/reddit.ts"],
-  "sourcesContent": ["import { z } from 'zod';\n\n// ============================================================================\n// search-reddit \u2014 input schema\n// ============================================================================\n\nexport const searchRedditParamsSchema = z.object({\n  queries: z\n    .array(\n      z\n        .string()\n        .min(1, { message: 'search-reddit: Query cannot be empty' })\n        .describe('A Reddit search query. Do NOT add \"site:reddit.com\" \u2014 it is appended automatically.'),\n    )\n    .min(1, { message: 'search-reddit: At least 1 query is required' })\n    .max(50, { message: 'search-reddit: Maximum 50 queries allowed' })\n    .describe('Array of 1-50 search queries. Each query gets \"site:reddit.com\" appended and is sent to Serper as a standard Google search. Returns a flat list of deduplicated Reddit URLs. Use with get-reddit-post to fetch full content and extract insights.'),\n}).strict();\n\nexport type SearchRedditParams = z.infer<typeof searchRedditParamsSchema>;\n\n// ============================================================================\n// get-reddit-post \u2014 input schema\n// ============================================================================\n\nexport const getRedditPostParamsSchema = z.object({\n  urls: z\n    .array(\n      z\n        .string()\n        .url({ message: 'get-reddit-post: Each URL must be valid' })\n        .describe('A full Reddit post URL (e.g., \"https://www.reddit.com/r/subreddit/comments/id/title/\").'),\n    )\n    .min(1, { message: 'get-reddit-post: At least 1 Reddit post URL is required' })\n    .max(50, { message: 'get-reddit-post: Maximum 50 Reddit post URLs allowed' })\n    .describe('Array of 1-50 Reddit post URLs. Each post is fetched with full comment trees, then the LLM extracts insights per what_to_extract. Best used after search-reddit.'),\n  fetch_comments: z\n    .boolean()\n    .default(true)\n    .describe('Fetch threaded comment trees for each post. Defaults to true. Comments include author, score, OP markers, and nested replies. Set false only when you need post titles/selftext without community discussion.'),\n  what_to_extract: z\n    .string({ error: 'get-reddit-post: what_to_extract is required' })\n    .min(5, { message: 'get-reddit-post: what_to_extract must be at least 5 characters' })\n    .max(1000, { message: 'get-reddit-post: what_to_extract is too long (max 1000 characters)' })\n    .describe('REQUIRED. Extraction instructions for the LLM. Describes what insights, opinions, or data to pull from each post and its comments. Use pipe separators for multiple targets: \"Extract recommendations | pain points | consensus on best practices | specific tools mentioned\".'),\n}).strict();\n\nexport type GetRedditPostParams = z.infer<typeof getRedditPostParamsSchema>;\n\n// ============================================================================\n// search-reddit \u2014 output schema\n// ============================================================================\n\nexport const searchRedditOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('Newline-separated list of unique Reddit URLs discovered across all queries.'),\n  metadata: z.object({\n    query_count: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Number of queries executed.'),\n    total_urls: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total unique Reddit URLs returned.'),\n  }).strict().describe('Metadata about the Reddit URL search.'),\n}).strict();\n\nexport type SearchRedditOutput = z.infer<typeof searchRedditOutputSchema>;\n\n// ============================================================================\n// get-reddit-post \u2014 output schema\n// ============================================================================\n\nexport const getRedditPostOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('LLM-synthesized extraction from Reddit posts and comments, structured per what_to_extract instructions.'),\n  metadata: z.object({\n    total_urls: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total number of Reddit post URLs processed.'),\n    successful: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Number of posts fetched successfully.'),\n    failed: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Number of post fetches that failed.'),\n    fetch_comments: z\n      .boolean()\n      .describe('Whether comments were fetched for each post.'),\n    total_words_used: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total words used across all posts.'),\n    llm_failures: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Count of posts where LLM extraction failed (raw content returned instead).'),\n    total_batches: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Number of Reddit API batches executed.'),\n    rate_limit_hits: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Observed Reddit API rate-limit retries during the batch.'),\n  }).strict().describe('Metadata about the Reddit post fetch and extraction.'),\n}).strict();\n\nexport type GetRedditPostOutput = z.infer<typeof getRedditPostOutputSchema>;\n"],
-  "mappings": "AAAA,SAAS,SAAS;AAMX,MAAM,2BAA2B,EAAE,OAAO;AAAA,EAC/C,SAAS,EACN;AAAA,IACC,EACG,OAAO,EACP,IAAI,GAAG,EAAE,SAAS,uCAAuC,CAAC,EAC1D,SAAS,0FAAqF;AAAA,EACnG,EACC,IAAI,GAAG,EAAE,SAAS,8CAA8C,CAAC,EACjE,IAAI,IAAI,EAAE,SAAS,4CAA4C,CAAC,EAChE,SAAS,mPAAmP;AACjQ,CAAC,EAAE,OAAO;AAQH,MAAM,4BAA4B,EAAE,OAAO;AAAA,EAChD,MAAM,EACH;AAAA,IACC,EACG,OAAO,EACP,IAAI,EAAE,SAAS,0CAA0C,CAAC,EAC1D,SAAS,yFAAyF;AAAA,EACvG,EACC,IAAI,GAAG,EAAE,SAAS,0DAA0D,CAAC,EAC7E,IAAI,IAAI,EAAE,SAAS,uDAAuD,CAAC,EAC3E,SAAS,kKAAkK;AAAA,EAC9K,gBAAgB,EACb,QAAQ,EACR,QAAQ,IAAI,EACZ,SAAS,+MAA+M;AAAA,EAC3N,iBAAiB,EACd,OAAO,EAAE,OAAO,+CAA+C,CAAC,EAChE,IAAI,GAAG,EAAE,SAAS,iEAAiE,CAAC,EACpF,IAAI,KAAM,EAAE,SAAS,qEAAqE,CAAC,EAC3F,SAAS,gRAAgR;AAC9R,CAAC,EAAE,OAAO;AAQH,MAAM,2BAA2B,EAAE,OAAO;AAAA,EAC/C,SAAS,EACN,OAAO,EACP,SAAS,6EAA6E;AAAA,EACzF,UAAU,EAAE,OAAO;AAAA,IACjB,aAAa,EACV,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,6BAA6B;AAAA,IACzC,YAAY,EACT,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,oCAAoC;AAAA,EAClD,CAAC,EAAE,OAAO,EAAE,SAAS,uCAAuC;AAC9D,CAAC,EAAE,OAAO;AAQH,MAAM,4BAA4B,EAAE,OAAO;AAAA,EAChD,SAAS,EACN,OAAO,EACP,SAAS,yGAAyG;AAAA,EACrH,UAAU,EAAE,OAAO;AAAA,IACjB,YAAY,EACT,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,6CAA6C;AAAA,IACzD,YAAY,EACT,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,uCAAuC;AAAA,IACnD,QAAQ,EACL,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,qCAAqC;AAAA,IACjD,gBAAgB,EACb,QAAQ,EACR,SAAS,8CAA8C;AAAA,IAC1D,kBAAkB,EACf,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,oCAAoC;AAAA,IAChD,cAAc,EACX,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,4EAA4E;AAAA,IACxF,eAAe,EACZ,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,wCAAwC;AAAA,IACpD,iBAAiB,EACd,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,0DAA0D;AAAA,EACxE,CAAC,EAAE,OAAO,EAAE,SAAS,sDAAsD;AAC7E,CAAC,EAAE,OAAO;",
+  "sourcesContent": ["import { z } from 'zod';\n\n// ============================================================================\n// search-reddit\n// ============================================================================\n\nexport const searchRedditParamsSchema = z.object({\n  queries: z\n    .array(\n      z.string()\n        .min(1, { message: 'search-reddit: Query cannot be empty' })\n        .describe('A Reddit search query. \"site:reddit.com\" is appended automatically.'),\n    )\n    .min(1, { message: 'search-reddit: At least 1 query required' })\n    .max(100, { message: 'search-reddit: Maximum 100 queries allowed' })\n    .describe('Search queries for Reddit. Each query is automatically scoped to reddit.com via Google. Returns deduplicated Reddit post URLs.'),\n}).strict();\n\nexport type SearchRedditParams = z.infer<typeof searchRedditParamsSchema>;\n\nexport const searchRedditOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('Newline-separated list of unique Reddit URLs.'),\n  metadata: z.object({\n    total_items: z.number().int().nonnegative().describe('Number of queries executed.'),\n    successful: z.number().int().nonnegative().describe('Queries that returned results.'),\n    failed: z.number().int().nonnegative().describe('Queries that failed.'),\n    execution_time_ms: z.number().int().nonnegative().describe('Wall clock time in milliseconds.'),\n  }).strict(),\n}).strict();\n\nexport type SearchRedditOutput = z.infer<typeof searchRedditOutputSchema>;\n\n// ============================================================================\n// get-reddit-post\n// ============================================================================\n\nexport const getRedditPostParamsSchema = z.object({\n  urls: z\n    .array(\n      z.string()\n        .url({ message: 'get-reddit-post: Each URL must be valid' })\n        .describe('A Reddit post URL.'),\n    )\n    .min(1, { message: 'get-reddit-post: At least 1 URL required' })\n    .max(100, { message: 'get-reddit-post: Maximum 100 URLs allowed' })\n    .describe('Reddit post URLs to fetch. Each post is returned with its full threaded comment tree.'),\n}).strict();\n\nexport type GetRedditPostParams = z.infer<typeof getRedditPostParamsSchema>;\n\nexport const getRedditPostOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('Raw Reddit posts with threaded comments including author, score, and OP markers.'),\n  metadata: z.object({\n    total_items: z.number().int().nonnegative().describe('Number of URLs processed.'),\n    successful: z.number().int().nonnegative().describe('Posts fetched successfully.'),\n    failed: z.number().int().nonnegative().describe('Posts that failed to fetch.'),\n    execution_time_ms: z.number().int().nonnegative().describe('Wall clock time in milliseconds.'),\n    rate_limit_hits: z.number().int().nonnegative().describe('Reddit API rate-limit retries.'),\n  }).strict(),\n}).strict();\n\nexport type GetRedditPostOutput = z.infer<typeof getRedditPostOutputSchema>;\n"],
+  "mappings": "AAAA,SAAS,SAAS;AAMX,MAAM,2BAA2B,EAAE,OAAO;AAAA,EAC/C,SAAS,EACN;AAAA,IACC,EAAE,OAAO,EACN,IAAI,GAAG,EAAE,SAAS,uCAAuC,CAAC,EAC1D,SAAS,qEAAqE;AAAA,EACnF,EACC,IAAI,GAAG,EAAE,SAAS,2CAA2C,CAAC,EAC9D,IAAI,KAAK,EAAE,SAAS,6CAA6C,CAAC,EAClE,SAAS,gIAAgI;AAC9I,CAAC,EAAE,OAAO;AAIH,MAAM,2BAA2B,EAAE,OAAO;AAAA,EAC/C,SAAS,EACN,OAAO,EACP,SAAS,+CAA+C;AAAA,EAC3D,UAAU,EAAE,OAAO;AAAA,IACjB,aAAa,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,6BAA6B;AAAA,IAClF,YAAY,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,gCAAgC;AAAA,IACpF,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,sBAAsB;AAAA,IACtE,mBAAmB,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,kCAAkC;AAAA,EAC/F,CAAC,EAAE,OAAO;AACZ,CAAC,EAAE,OAAO;AAQH,MAAM,4BAA4B,EAAE,OAAO;AAAA,EAChD,MAAM,EACH;AAAA,IACC,EAAE,OAAO,EACN,IAAI,EAAE,SAAS,0CAA0C,CAAC,EAC1D,SAAS,oBAAoB;AAAA,EAClC,EACC,IAAI,GAAG,EAAE,SAAS,2CAA2C,CAAC,EAC9D,IAAI,KAAK,EAAE,SAAS,4CAA4C,CAAC,EACjE,SAAS,uFAAuF;AACrG,CAAC,EAAE,OAAO;AAIH,MAAM,4BAA4B,EAAE,OAAO;AAAA,EAChD,SAAS,EACN,OAAO,EACP,SAAS,kFAAkF;AAAA,EAC9F,UAAU,EAAE,OAAO;AAAA,IACjB,aAAa,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,2BAA2B;AAAA,IAChF,YAAY,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,6BAA6B;AAAA,IACjF,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,6BAA6B;AAAA,IAC7E,mBAAmB,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,kCAAkC;AAAA,IAC7F,iBAAiB,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,gCAAgC;AAAA,EAC3F,CAAC,EAAE,OAAO;AACZ,CAAC,EAAE,OAAO;",
   "names": []
 }

package/dist/src/schemas/scrape-links.js

@@ -1,25 +1,21 @@
 import { z } from "zod";
-const urlSchema = z.string({ error: "scrape-links: URL is required" }).url({ message: "scrape-links: Invalid URL format" }).refine(
+const urlSchema = z.string().url({ message: "scrape-links: Invalid URL format" }).refine(
   (url) => url.startsWith("http://") || url.startsWith("https://"),
-  { message: "scrape-links: URL must use http:// or https:// protocol" }
-).describe("A fully-qualified HTTP or HTTPS URL to fetch and extract content from.");
+  { message: "scrape-links: URL must use http:// or https://" }
+).describe("A fully-qualified HTTP or HTTPS URL to scrape.");
 const scrapeLinksParamsSchema = z.object({
-  urls: z.array(urlSchema, { error: "scrape-links: URLs must be an array" }).min(1, { message: "scrape-links: At least 1 URL is required" }).max(50, { message: "scrape-links: Maximum 50 URLs allowed per request" }).describe("URLs to scrape (1-50). Token budget (32K) is split across URLs: 3 URLs get ~10K tokens each (deep), 10 get ~3K (balanced), 50 get ~640 (scan). Each page is scraped, cleaned, and processed by the LLM per what_to_extract."),
-  timeout: z.number({ error: "scrape-links: Timeout must be a number" }).min(5, { message: "scrape-links: Timeout must be at least 5 seconds" }).max(120, { message: "scrape-links: Timeout cannot exceed 120 seconds" }).default(30).describe("Timeout in seconds for each URL."),
-  what_to_extract: z.string({ error: "scrape-links: what_to_extract is required" }).min(5, { message: "scrape-links: what_to_extract must be at least 5 characters" }).max(1e3, { message: "scrape-links: Extraction instructions too long (max 1000 characters)" }).describe('REQUIRED. Extraction instructions for the LLM. The LLM processes each scraped page and extracts ONLY what you specify. Formula: "Extract [target1] | [target2] | [target3] with focus on [aspect]". Be specific: "pricing tiers | monthly vs annual cost | free tier limits" not just "pricing".')
+  urls: z.array(urlSchema).min(1, { message: "scrape-links: At least 1 URL required" }).max(100, { message: "scrape-links: Maximum 100 URLs allowed" }).describe("Web page URLs to scrape and extract content from."),
+  extract: z.string().min(5, { message: "scrape-links: extract must be at least 5 characters" }).max(1e3, { message: "scrape-links: extract too long (max 1000 chars)" }).describe('What to pull from each page. The LLM reads the scraped content and returns only what you specify. Be specific: "pricing tiers | free tier limits | enterprise contact info" not "pricing".')
 }).strict();
 const scrapeLinksOutputSchema = z.object({
-  content: z.string().describe("LLM-extracted content from scraped pages, structured per what_to_extract instructions."),
+  content: z.string().describe("LLM-extracted content from scraped pages per the extract instructions."),
   metadata: z.object({
-    total_urls: z.number().int().nonnegative().describe("Total number of input URLs processed."),
-    successful: z.number().int().nonnegative().describe("Number of URLs that were fetched successfully."),
-    failed: z.number().int().nonnegative().describe("Number of URLs that failed validation or scraping."),
-    total_credits: z.number().int().nonnegative().describe("Total external scraping credits consumed."),
-    execution_time_ms: z.number().int().nonnegative().describe("Elapsed execution time in milliseconds."),
-    tokens_per_url: z.number().int().nonnegative().optional().describe("Allocated LLM token budget per successfully scraped URL."),
-    total_token_budget: z.number().int().nonnegative().optional().describe("Overall token budget available for extraction."),
-    batches_processed: z.number().int().nonnegative().optional().describe("Number of scrape batches executed.")
-  }).strict().describe("Structured metadata about the scrape and extraction batch.")
+    total_items: z.number().int().nonnegative().describe("Number of URLs processed."),
+    successful: z.number().int().nonnegative().describe("URLs fetched successfully."),
+    failed: z.number().int().nonnegative().describe("URLs that failed."),
+    execution_time_ms: z.number().int().nonnegative().describe("Wall clock time in milliseconds."),
+    total_credits: z.number().int().nonnegative().describe("External scraping credits consumed.")
+  }).strict()
 }).strict();
 export {
   scrapeLinksOutputSchema,

package/dist/src/schemas/scrape-links.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../src/schemas/scrape-links.ts"],
-  "sourcesContent": ["import { z } from 'zod';\n\n// URL schema with protocol validation\nconst urlSchema = z\n  .string({ error: 'scrape-links: URL is required' })\n  .url({ message: 'scrape-links: Invalid URL format' })\n  .refine(\n    url => url.startsWith('http://') || url.startsWith('https://'),\n    { message: 'scrape-links: URL must use http:// or https:// protocol' }\n  )\n  .describe('A fully-qualified HTTP or HTTPS URL to fetch and extract content from.');\n\n// Input schema for scrape-links tool\nexport const scrapeLinksParamsSchema = z.object({\n  urls: z\n    .array(urlSchema, { error: 'scrape-links: URLs must be an array' })\n    .min(1, { message: 'scrape-links: At least 1 URL is required' })\n    .max(50, { message: 'scrape-links: Maximum 50 URLs allowed per request' })\n    .describe('URLs to scrape (1-50). Token budget (32K) is split across URLs: 3 URLs get ~10K tokens each (deep), 10 get ~3K (balanced), 50 get ~640 (scan). Each page is scraped, cleaned, and processed by the LLM per what_to_extract.'),\n  timeout: z\n    .number({ error: 'scrape-links: Timeout must be a number' })\n    .min(5, { message: 'scrape-links: Timeout must be at least 5 seconds' })\n    .max(120, { message: 'scrape-links: Timeout cannot exceed 120 seconds' })\n    .default(30)\n    .describe('Timeout in seconds for each URL.'),\n  what_to_extract: z\n    .string({ error: 'scrape-links: what_to_extract is required' })\n    .min(5, { message: 'scrape-links: what_to_extract must be at least 5 characters' })\n    .max(1000, { message: 'scrape-links: Extraction instructions too long (max 1000 characters)' })\n    .describe('REQUIRED. Extraction instructions for the LLM. The LLM processes each scraped page and extracts ONLY what you specify. Formula: \"Extract [target1] | [target2] | [target3] with focus on [aspect]\". Be specific: \"pricing tiers | monthly vs annual cost | free tier limits\" not just \"pricing\".'),\n}).strict();\n\nexport type ScrapeLinksParams = z.infer<typeof scrapeLinksParamsSchema>;\n\nexport const scrapeLinksOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('LLM-extracted content from scraped pages, structured per what_to_extract instructions.'),\n  metadata: z.object({\n    total_urls: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total number of input URLs processed.'),\n    successful: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Number of URLs that were fetched successfully.'),\n    failed: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Number of URLs that failed validation or scraping.'),\n    total_credits: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total external scraping credits consumed.'),\n    execution_time_ms: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Elapsed execution time in milliseconds.'),\n    tokens_per_url: z\n      .number()\n      .int()\n      .nonnegative()\n      .optional()\n      .describe('Allocated LLM token budget per successfully scraped URL.'),\n    total_token_budget: z\n      .number()\n      .int()\n      .nonnegative()\n      .optional()\n      .describe('Overall token budget available for extraction.'),\n    batches_processed: z\n      .number()\n      .int()\n      .nonnegative()\n      .optional()\n      .describe('Number of scrape batches executed.'),\n  }).strict().describe('Structured metadata about the scrape and extraction batch.'),\n}).strict();\n\nexport type ScrapeLinksOutput = z.infer<typeof scrapeLinksOutputSchema>;\n"],
-  "mappings": "AAAA,SAAS,SAAS;AAGlB,MAAM,YAAY,EACf,OAAO,EAAE,OAAO,gCAAgC,CAAC,EACjD,IAAI,EAAE,SAAS,mCAAmC,CAAC,EACnD;AAAA,EACC,SAAO,IAAI,WAAW,SAAS,KAAK,IAAI,WAAW,UAAU;AAAA,EAC7D,EAAE,SAAS,0DAA0D;AACvE,EACC,SAAS,wEAAwE;AAG7E,MAAM,0BAA0B,EAAE,OAAO;AAAA,EAC9C,MAAM,EACH,MAAM,WAAW,EAAE,OAAO,sCAAsC,CAAC,EACjE,IAAI,GAAG,EAAE,SAAS,2CAA2C,CAAC,EAC9D,IAAI,IAAI,EAAE,SAAS,oDAAoD,CAAC,EACxE,SAAS,6NAA6N;AAAA,EACzO,SAAS,EACN,OAAO,EAAE,OAAO,yCAAyC,CAAC,EAC1D,IAAI,GAAG,EAAE,SAAS,mDAAmD,CAAC,EACtE,IAAI,KAAK,EAAE,SAAS,kDAAkD,CAAC,EACvE,QAAQ,EAAE,EACV,SAAS,kCAAkC;AAAA,EAC9C,iBAAiB,EACd,OAAO,EAAE,OAAO,4CAA4C,CAAC,EAC7D,IAAI,GAAG,EAAE,SAAS,8DAA8D,CAAC,EACjF,IAAI,KAAM,EAAE,SAAS,uEAAuE,CAAC,EAC7F,SAAS,kSAAkS;AAChT,CAAC,EAAE,OAAO;AAIH,MAAM,0BAA0B,EAAE,OAAO;AAAA,EAC9C,SAAS,EACN,OAAO,EACP,SAAS,wFAAwF;AAAA,EACpG,UAAU,EAAE,OAAO;AAAA,IACjB,YAAY,EACT,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,uCAAuC;AAAA,IACnD,YAAY,EACT,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,gDAAgD;AAAA,IAC5D,QAAQ,EACL,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,oDAAoD;AAAA,IAChE,eAAe,EACZ,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,2CAA2C;AAAA,IACvD,mBAAmB,EAChB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,yCAAyC;AAAA,IACrD,gBAAgB,EACb,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,EACT,SAAS,0DAA0D;AAAA,IACtE,oBAAoB,EACjB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,EACT,SAAS,gDAAgD;AAAA,IAC5D,mBAAmB,EAChB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,EACT,SAAS,oCAAoC;AAAA,EAClD,CAAC,EAAE,OAAO,EAAE,SAAS,4DAA4D;AACnF,CAAC,EAAE,OAAO;",
+  "sourcesContent": ["import { z } from 'zod';\n\nconst urlSchema = z\n  .string()\n  .url({ message: 'scrape-links: Invalid URL format' })\n  .refine(\n    url => url.startsWith('http://') || url.startsWith('https://'),\n    { message: 'scrape-links: URL must use http:// or https://' }\n  )\n  .describe('A fully-qualified HTTP or HTTPS URL to scrape.');\n\nexport const scrapeLinksParamsSchema = z.object({\n  urls: z\n    .array(urlSchema)\n    .min(1, { message: 'scrape-links: At least 1 URL required' })\n    .max(100, { message: 'scrape-links: Maximum 100 URLs allowed' })\n    .describe('Web page URLs to scrape and extract content from.'),\n  extract: z\n    .string()\n    .min(5, { message: 'scrape-links: extract must be at least 5 characters' })\n    .max(1000, { message: 'scrape-links: extract too long (max 1000 chars)' })\n    .describe('What to pull from each page. The LLM reads the scraped content and returns only what you specify. Be specific: \"pricing tiers | free tier limits | enterprise contact info\" not \"pricing\".'),\n}).strict();\n\nexport type ScrapeLinksParams = z.infer<typeof scrapeLinksParamsSchema>;\n\nexport const scrapeLinksOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('LLM-extracted content from scraped pages per the extract instructions.'),\n  metadata: z.object({\n    total_items: z.number().int().nonnegative().describe('Number of URLs processed.'),\n    successful: z.number().int().nonnegative().describe('URLs fetched successfully.'),\n    failed: z.number().int().nonnegative().describe('URLs that failed.'),\n    execution_time_ms: z.number().int().nonnegative().describe('Wall clock time in milliseconds.'),\n    total_credits: z.number().int().nonnegative().describe('External scraping credits consumed.'),\n  }).strict(),\n}).strict();\n\nexport type ScrapeLinksOutput = z.infer<typeof scrapeLinksOutputSchema>;\n"],
+  "mappings": "AAAA,SAAS,SAAS;AAElB,MAAM,YAAY,EACf,OAAO,EACP,IAAI,EAAE,SAAS,mCAAmC,CAAC,EACnD;AAAA,EACC,SAAO,IAAI,WAAW,SAAS,KAAK,IAAI,WAAW,UAAU;AAAA,EAC7D,EAAE,SAAS,iDAAiD;AAC9D,EACC,SAAS,gDAAgD;AAErD,MAAM,0BAA0B,EAAE,OAAO;AAAA,EAC9C,MAAM,EACH,MAAM,SAAS,EACf,IAAI,GAAG,EAAE,SAAS,wCAAwC,CAAC,EAC3D,IAAI,KAAK,EAAE,SAAS,yCAAyC,CAAC,EAC9D,SAAS,mDAAmD;AAAA,EAC/D,SAAS,EACN,OAAO,EACP,IAAI,GAAG,EAAE,SAAS,sDAAsD,CAAC,EACzE,IAAI,KAAM,EAAE,SAAS,kDAAkD,CAAC,EACxE,SAAS,4LAA4L;AAC1M,CAAC,EAAE,OAAO;AAIH,MAAM,0BAA0B,EAAE,OAAO;AAAA,EAC9C,SAAS,EACN,OAAO,EACP,SAAS,wEAAwE;AAAA,EACpF,UAAU,EAAE,OAAO;AAAA,IACjB,aAAa,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,2BAA2B;AAAA,IAChF,YAAY,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,4BAA4B;AAAA,IAChF,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,mBAAmB;AAAA,IACnE,mBAAmB,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,kCAAkC;AAAA,IAC7F,eAAe,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,qCAAqC;AAAA,EAC9F,CAAC,EAAE,OAAO;AACZ,CAAC,EAAE,OAAO;",
   "names": []
 }

package/dist/src/schemas/web-search.js

@@ -1,31 +1,27 @@
 import { z } from "zod";
-const keywordSchema = z.string({ error: "web-search: Keyword is required" }).min(1, { message: "web-search: Keyword cannot be empty" }).max(500, { message: "web-search: Keyword too long (max 500 characters)" }).refine(
-  (k) => k.trim().length > 0,
-  { message: "web-search: Keyword cannot be whitespace only" }
-).describe('A single Google search query (1-500 chars). Each keyword runs as a separate parallel search. Use varied angles: direct topic, comparisons, "best of" lists, year-specific, site-specific (e.g., "site:github.com topic").');
 const webSearchParamsSchema = z.object({
-  keywords: z.array(keywordSchema, { error: "web-search: Keywords must be an array" }).min(1, { message: "web-search: At least 1 keyword required" }).max(100, { message: "web-search: Maximum 100 keywords allowed per request" }).describe("Array of 1-100 search keywords. Each runs as a separate Google search in parallel. Results are aggregated, deduplicated, and ranked by CTR-weighted consensus. RECOMMENDED: 3-7 keywords for solid consensus, up to 20 for thorough coverage."),
-  objective: z.string({ error: "web-search: objective is required" }).min(5, { message: "web-search: objective must be at least 5 characters" }).max(500, { message: "web-search: objective too long (max 500 characters)" }).describe('REQUIRED. Describes what you are looking for. An LLM classifies each search result into 3 relevance tiers (highly relevant, maybe relevant, other) using only titles, snippets, and site names \u2014 no URLs are fetched. Be specific: "open-source MCP server implementations in TypeScript" not "MCP servers". Also generates a synthesis paragraph summarizing key findings.'),
-  raw: z.boolean({ error: "web-search: raw must be a boolean" }).default(false).describe("When true, skip LLM classification and return the traditional CTR-weighted consensus-ranked URL list. Use when you need raw context or the LLM endpoint is unavailable. Default: false (LLM classification enabled).")
+  queries: z.array(
+    z.string().min(1, { message: "web-search: Query cannot be empty" }).max(500, { message: "web-search: Query too long (max 500 chars)" }).describe("A single Google search query. Each query runs as a separate parallel search.")
+  ).min(1, { message: "web-search: At least 1 query required" }).max(100, { message: "web-search: Maximum 100 queries allowed" }).describe("Search queries to run in parallel via Google. More queries = broader coverage and stronger consensus signals across results."),
+  extract: z.string().min(5, { message: "web-search: extract must be at least 5 characters" }).max(500, { message: "web-search: extract too long (max 500 chars)" }).describe('What you are looking for. The LLM classifies each result by relevance and generates a synthesis. Be specific: "TypeScript MCP server frameworks with OAuth support" not "MCP servers".'),
+  raw: z.boolean().default(false).describe("Skip LLM classification and return the raw ranked URL list. Use when you need unprocessed results.")
 }).strict();
 const webSearchOutputSchema = z.object({
-  content: z.string().describe("Markdown report. With LLM: 3-tier table (highly relevant / maybe relevant / other) with synthesis. With raw=true: traditional CTR-ranked list."),
+  content: z.string().describe("Markdown report with tiered results (LLM mode) or ranked URL list (raw mode)."),
   metadata: z.object({
-    total_keywords: z.number().int().nonnegative().describe("Total number of keyword queries executed."),
-    total_results: z.number().int().nonnegative().describe("Total unique URLs found across all searches."),
-    execution_time_ms: z.number().int().nonnegative().describe("Elapsed execution time in milliseconds."),
-    total_unique_urls: z.number().int().nonnegative().optional().describe("Unique URL count observed across all searches."),
-    consensus_url_count: z.number().int().nonnegative().optional().describe("Count of URLs that met the consensus threshold."),
-    frequency_threshold: z.number().int().nonnegative().optional().describe("Minimum frequency required for a URL to be considered consensus."),
-    llm_classified: z.boolean().describe("Whether LLM classification was applied to the results."),
-    llm_error: z.string().optional().describe("LLM classification error message if classification failed and fell back to raw output."),
+    total_items: z.number().int().nonnegative().describe("Number of queries executed."),
+    successful: z.number().int().nonnegative().describe("Queries that returned results."),
+    failed: z.number().int().nonnegative().describe("Queries that failed."),
+    execution_time_ms: z.number().int().nonnegative().describe("Wall clock time in milliseconds."),
+    llm_classified: z.boolean().describe("Whether LLM classification was applied."),
+    llm_error: z.string().optional().describe("LLM error if classification failed and fell back to raw."),
     coverage_summary: z.array(z.object({
-      keyword: z.string().describe("The search keyword."),
-      result_count: z.number().int().nonnegative().describe("Number of results returned for this keyword."),
-      top_url: z.string().optional().describe("Domain of the top-ranked result for this keyword.")
-    })).optional().describe("Per-keyword result counts and top URLs for coverage analysis."),
-    low_yield_keywords: z.array(z.string()).optional().describe("Keywords that produced 0-1 results.")
-  }).strict().describe("Structured metadata about the completed web search batch.")
+      keyword: z.string().describe("The search query."),
+      result_count: z.number().int().nonnegative().describe("Results returned for this query."),
+      top_url: z.string().optional().describe("Domain of the top result.")
+    })).optional().describe("Per-query result counts and top URLs."),
+    low_yield_keywords: z.array(z.string()).optional().describe("Queries that produced 0-1 results.")
+  }).strict()
 }).strict();
 export {
   webSearchOutputSchema,

package/dist/src/schemas/web-search.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../src/schemas/web-search.ts"],
-  "sourcesContent": ["import { z } from 'zod';\n\n// Keyword schema with validation\nconst keywordSchema = z\n  .string({ error: 'web-search: Keyword is required' })\n  .min(1, { message: 'web-search: Keyword cannot be empty' })\n  .max(500, { message: 'web-search: Keyword too long (max 500 characters)' })\n  .refine(\n    k => k.trim().length > 0,\n    { message: 'web-search: Keyword cannot be whitespace only' }\n  )\n  .describe('A single Google search query (1-500 chars). Each keyword runs as a separate parallel search. Use varied angles: direct topic, comparisons, \"best of\" lists, year-specific, site-specific (e.g., \"site:github.com topic\").');\n\n// Input schema for web-search tool\nexport const webSearchParamsSchema = z.object({\n  keywords: z\n    .array(keywordSchema, { error: 'web-search: Keywords must be an array' })\n    .min(1, { message: 'web-search: At least 1 keyword required' })\n    .max(100, { message: 'web-search: Maximum 100 keywords allowed per request' })\n    .describe('Array of 1-100 search keywords. Each runs as a separate Google search in parallel. Results are aggregated, deduplicated, and ranked by CTR-weighted consensus. RECOMMENDED: 3-7 keywords for solid consensus, up to 20 for thorough coverage.'),\n  objective: z\n    .string({ error: 'web-search: objective is required' })\n    .min(5, { message: 'web-search: objective must be at least 5 characters' })\n    .max(500, { message: 'web-search: objective too long (max 500 characters)' })\n    .describe('REQUIRED. Describes what you are looking for. An LLM classifies each search result into 3 relevance tiers (highly relevant, maybe relevant, other) using only titles, snippets, and site names \u2014 no URLs are fetched. Be specific: \"open-source MCP server implementations in TypeScript\" not \"MCP servers\". Also generates a synthesis paragraph summarizing key findings.'),\n  raw: z\n    .boolean({ error: 'web-search: raw must be a boolean' })\n    .default(false)\n    .describe('When true, skip LLM classification and return the traditional CTR-weighted consensus-ranked URL list. Use when you need raw context or the LLM endpoint is unavailable. Default: false (LLM classification enabled).'),\n}).strict();\n\nexport type WebSearchParams = z.infer<typeof webSearchParamsSchema>;\n\nexport const webSearchOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('Markdown report. With LLM: 3-tier table (highly relevant / maybe relevant / other) with synthesis. With raw=true: traditional CTR-ranked list.'),\n  metadata: z.object({\n    total_keywords: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total number of keyword queries executed.'),\n    total_results: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Total unique URLs found across all searches.'),\n    execution_time_ms: z\n      .number()\n      .int()\n      .nonnegative()\n      .describe('Elapsed execution time in milliseconds.'),\n    total_unique_urls: z\n      .number()\n      .int()\n      .nonnegative()\n      .optional()\n      .describe('Unique URL count observed across all searches.'),\n    consensus_url_count: z\n      .number()\n      .int()\n      .nonnegative()\n      .optional()\n      .describe('Count of URLs that met the consensus threshold.'),\n    frequency_threshold: z\n      .number()\n      .int()\n      .nonnegative()\n      .optional()\n      .describe('Minimum frequency required for a URL to be considered consensus.'),\n    llm_classified: z\n      .boolean()\n      .describe('Whether LLM classification was applied to the results.'),\n    llm_error: z\n      .string()\n      .optional()\n      .describe('LLM classification error message if classification failed and fell back to raw output.'),\n    coverage_summary: z\n      .array(z.object({\n        keyword: z.string().describe('The search keyword.'),\n        result_count: z.number().int().nonnegative().describe('Number of results returned for this keyword.'),\n        top_url: z.string().optional().describe('Domain of the top-ranked result for this keyword.'),\n      }))\n      .optional()\n      .describe('Per-keyword result counts and top URLs for coverage analysis.'),\n    low_yield_keywords: z\n      .array(z.string())\n      .optional()\n      .describe('Keywords that produced 0-1 results.'),\n  }).strict().describe('Structured metadata about the completed web search batch.'),\n}).strict();\n\nexport type WebSearchOutput = z.infer<typeof webSearchOutputSchema>;\n"],
-  "mappings": "AAAA,SAAS,SAAS;AAGlB,MAAM,gBAAgB,EACnB,OAAO,EAAE,OAAO,kCAAkC,CAAC,EACnD,IAAI,GAAG,EAAE,SAAS,sCAAsC,CAAC,EACzD,IAAI,KAAK,EAAE,SAAS,oDAAoD,CAAC,EACzE;AAAA,EACC,OAAK,EAAE,KAAK,EAAE,SAAS;AAAA,EACvB,EAAE,SAAS,gDAAgD;AAC7D,EACC,SAAS,2NAA2N;AAGhO,MAAM,wBAAwB,EAAE,OAAO;AAAA,EAC5C,UAAU,EACP,MAAM,eAAe,EAAE,OAAO,wCAAwC,CAAC,EACvE,IAAI,GAAG,EAAE,SAAS,0CAA0C,CAAC,EAC7D,IAAI,KAAK,EAAE,SAAS,uDAAuD,CAAC,EAC5E,SAAS,+OAA+O;AAAA,EAC3P,WAAW,EACR,OAAO,EAAE,OAAO,oCAAoC,CAAC,EACrD,IAAI,GAAG,EAAE,SAAS,sDAAsD,CAAC,EACzE,IAAI,KAAK,EAAE,SAAS,sDAAsD,CAAC,EAC3E,SAAS,kXAA6W;AAAA,EACzX,KAAK,EACF,QAAQ,EAAE,OAAO,oCAAoC,CAAC,EACtD,QAAQ,KAAK,EACb,SAAS,sNAAsN;AACpO,CAAC,EAAE,OAAO;AAIH,MAAM,wBAAwB,EAAE,OAAO;AAAA,EAC5C,SAAS,EACN,OAAO,EACP,SAAS,gJAAgJ;AAAA,EAC5J,UAAU,EAAE,OAAO;AAAA,IACjB,gBAAgB,EACb,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,2CAA2C;AAAA,IACvD,eAAe,EACZ,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,8CAA8C;AAAA,IAC1D,mBAAmB,EAChB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,yCAAyC;AAAA,IACrD,mBAAmB,EAChB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,EACT,SAAS,gDAAgD;AAAA,IAC5D,qBAAqB,EAClB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,EACT,SAAS,iDAAiD;AAAA,IAC7D,qBAAqB,EAClB,OAAO,EACP,IAAI,EACJ,YAAY,EACZ,SAAS,EACT,SAAS,kEAAkE;AAAA,IAC9E,gBAAgB,EACb,QAAQ,EACR,SAAS,wDAAwD;AAAA,IACpE,WAAW,EACR,OAAO,EACP,SAAS,EACT,SAAS,wFAAwF;AAAA,IACpG,kBAAkB,EACf,MAAM,EAAE,OAAO;AAAA,MACd,SAAS,EAAE,OAAO,EAAE,SAAS,qBAAqB;AAAA,MAClD,cAAc,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,8CAA8C;AAAA,MACpG,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,SAAS,mDAAmD;AAAA,IAC7F,CAAC,CAAC,EACD,SAAS,EACT,SAAS,+DAA+D;AAAA,IAC3E,oBAAoB,EACjB,MAAM,EAAE,OAAO,CAAC,EAChB,SAAS,EACT,SAAS,qCAAqC;AAAA,EACnD,CAAC,EAAE,OAAO,EAAE,SAAS,2DAA2D;AAClF,CAAC,EAAE,OAAO;",
+  "sourcesContent": ["import { z } from 'zod';\n\nexport const webSearchParamsSchema = z.object({\n  queries: z\n    .array(\n      z.string()\n        .min(1, { message: 'web-search: Query cannot be empty' })\n        .max(500, { message: 'web-search: Query too long (max 500 chars)' })\n        .describe('A single Google search query. Each query runs as a separate parallel search.'),\n    )\n    .min(1, { message: 'web-search: At least 1 query required' })\n    .max(100, { message: 'web-search: Maximum 100 queries allowed' })\n    .describe('Search queries to run in parallel via Google. More queries = broader coverage and stronger consensus signals across results.'),\n  extract: z\n    .string()\n    .min(5, { message: 'web-search: extract must be at least 5 characters' })\n    .max(500, { message: 'web-search: extract too long (max 500 chars)' })\n    .describe('What you are looking for. The LLM classifies each result by relevance and generates a synthesis. Be specific: \"TypeScript MCP server frameworks with OAuth support\" not \"MCP servers\".'),\n  raw: z\n    .boolean()\n    .default(false)\n    .describe('Skip LLM classification and return the raw ranked URL list. Use when you need unprocessed results.'),\n}).strict();\n\nexport type WebSearchParams = z.infer<typeof webSearchParamsSchema>;\n\nexport const webSearchOutputSchema = z.object({\n  content: z\n    .string()\n    .describe('Markdown report with tiered results (LLM mode) or ranked URL list (raw mode).'),\n  metadata: z.object({\n    total_items: z.number().int().nonnegative().describe('Number of queries executed.'),\n    successful: z.number().int().nonnegative().describe('Queries that returned results.'),\n    failed: z.number().int().nonnegative().describe('Queries that failed.'),\n    execution_time_ms: z.number().int().nonnegative().describe('Wall clock time in milliseconds.'),\n    llm_classified: z.boolean().describe('Whether LLM classification was applied.'),\n    llm_error: z.string().optional().describe('LLM error if classification failed and fell back to raw.'),\n    coverage_summary: z\n      .array(z.object({\n        keyword: z.string().describe('The search query.'),\n        result_count: z.number().int().nonnegative().describe('Results returned for this query.'),\n        top_url: z.string().optional().describe('Domain of the top result.'),\n      }))\n      .optional()\n      .describe('Per-query result counts and top URLs.'),\n    low_yield_keywords: z\n      .array(z.string())\n      .optional()\n      .describe('Queries that produced 0-1 results.'),\n  }).strict(),\n}).strict();\n\nexport type WebSearchOutput = z.infer<typeof webSearchOutputSchema>;\n"],
+  "mappings": "AAAA,SAAS,SAAS;AAEX,MAAM,wBAAwB,EAAE,OAAO;AAAA,EAC5C,SAAS,EACN;AAAA,IACC,EAAE,OAAO,EACN,IAAI,GAAG,EAAE,SAAS,oCAAoC,CAAC,EACvD,IAAI,KAAK,EAAE,SAAS,6CAA6C,CAAC,EAClE,SAAS,8EAA8E;AAAA,EAC5F,EACC,IAAI,GAAG,EAAE,SAAS,wCAAwC,CAAC,EAC3D,IAAI,KAAK,EAAE,SAAS,0CAA0C,CAAC,EAC/D,SAAS,8HAA8H;AAAA,EAC1I,SAAS,EACN,OAAO,EACP,IAAI,GAAG,EAAE,SAAS,oDAAoD,CAAC,EACvE,IAAI,KAAK,EAAE,SAAS,+CAA+C,CAAC,EACpE,SAAS,wLAAwL;AAAA,EACpM,KAAK,EACF,QAAQ,EACR,QAAQ,KAAK,EACb,SAAS,oGAAoG;AAClH,CAAC,EAAE,OAAO;AAIH,MAAM,wBAAwB,EAAE,OAAO;AAAA,EAC5C,SAAS,EACN,OAAO,EACP,SAAS,+EAA+E;AAAA,EAC3F,UAAU,EAAE,OAAO;AAAA,IACjB,aAAa,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,6BAA6B;AAAA,IAClF,YAAY,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,gCAAgC;AAAA,IACpF,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,sBAAsB;AAAA,IACtE,mBAAmB,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,kCAAkC;AAAA,IAC7F,gBAAgB,EAAE,QAAQ,EAAE,SAAS,yCAAyC;AAAA,IAC9E,WAAW,EAAE,OAAO,EAAE,SAAS,EAAE,SAAS,0DAA0D;AAAA,IACpG,kBAAkB,EACf,MAAM,EAAE,OAAO;AAAA,MACd,SAAS,EAAE,OAAO,EAAE,SAAS,mBAAmB;AAAA,MAChD,cAAc,EAAE,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,SAAS,kCAAkC;AAAA,MACxF,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,SAAS,2BAA2B;AAAA,IACrE,CAAC,CAAC,EACD,SAAS,EACT,SAAS,uCAAuC;AAAA,IACnD,oBAAoB,EACjB,MAAM,EAAE,OAAO,CAAC,EAChB,SAAS,EACT,SAAS,oCAAoC;AAAA,EAClD,CAAC,EAAE,OAAO;AACZ,CAAC,EAAE,OAAO;",
   "names": []
 }

package/dist/src/tools/reddit.js

@@ -6,16 +6,12 @@ import {
 } from "../schemas/reddit.js";
 import { SearchClient } from "../clients/search.js";
 import { RedditClient } from "../clients/reddit.js";
-import { REDDIT, CONCURRENCY, getCapabilities, getMissingEnvMessage, parseEnv } from "../config/index.js";
+import { REDDIT, getCapabilities, getMissingEnvMessage, parseEnv } from "../config/index.js";
 import { classifyError } from "../utils/errors.js";
-import { createLLMProcessor, processContentWithLLM } from "../services/llm-processor.js";
-import { pMap } from "../utils/concurrency.js";
 import {
-  mcpLog,
   formatSuccess,
   formatError,
-  formatBatchHeader,
-  TOKEN_BUDGETS
+  formatBatchHeader
 } from "./utils.js";
 import {
   createToolReporter,
@@ -84,20 +80,17 @@ ${post.body}
     wordsUsed += commentsResult.wordsUsed;
     if (commentsResult.truncated > 0) {
       md += `
-_${commentsResult.truncated} more comments not shown (word budget reached). Use use_llm=true for AI-synthesized summary._
+_${commentsResult.truncated} more comments not shown (word budget reached)._
 
 `;
     }
-  } else if (!fetchComments) {
-    md += `_Comments not fetched (fetch_comments=false)_
-
-`;
   }
   return { md, wordsUsed, commentsShown, commentsTruncated };
 }
 async function handleSearchReddit(queries, apiKey, reporter = NOOP_REPORTER) {
   try {
-    const limited = queries.slice(0, 50);
+    const startTime = Date.now();
+    const limited = queries.slice(0, 100);
     const client = new SearchClient(apiKey);
     await reporter.log("info", `Searching Reddit with ${limited.length} queries`);
     await reporter.progress(15, 100, "Searching Reddit");
@@ -114,18 +107,21 @@ async function handleSearchReddit(queries, apiKey, reporter = NOOP_REPORTER) {
         message: `No Reddit URLs found for any of the ${limited.length} queries`,
         toolName: "search-reddit",
         howToFix: ["Try broader or simpler search terms", "Check spelling"],
-        alternatives: ['web-search(keywords=["topic reddit discussion"], objective="...") \u2014 broader Google search']
+        alternatives: ['web-search(queries=["topic reddit discussion"], extract="...") \u2014 broader Google search']
       }));
     }
     const urlList = [...allUrls];
     const content = urlList.join("\n");
     await reporter.log("info", `Found ${urlList.length} unique Reddit URLs across ${limited.length} queries`);
     await reporter.progress(100, 100, "Reddit search complete");
+    const executionTime = Date.now() - startTime;
     return toolSuccess(content, {
       content,
       metadata: {
-        query_count: limited.length,
-        total_urls: urlList.length
+        total_items: limited.length,
+        successful: urlList.length,
+        failed: 0,
+        execution_time_ms: executionTime
       }
     });
   } catch (error) {
@@ -139,11 +135,6 @@ async function handleSearchReddit(queries, apiKey, reporter = NOOP_REPORTER) {
     }));
   }
 }
-function enhanceExtractionInstruction(instruction) {
-  return `${instruction}
-
-${REDDIT.EXTRACTION_SUFFIX}`;
-}
 function validatePostCount(urlCount) {
   if (urlCount < REDDIT.MIN_POSTS) {
     return formatError({
@@ -166,35 +157,13 @@ function validatePostCount(urlCount) {
   }
   return null;
 }
-async function applyLlmToPost(postContent, result, url, llmProcessor, enhancedInstruction, tokensPerUrl, index, total) {
-  mcpLog("info", `[${index}/${total}] Applying LLM extraction to ${url}`, "reddit");
-  const llmResult = await processContentWithLLM(
-    postContent,
-    { use_llm: true, what_to_extract: enhancedInstruction, max_tokens: tokensPerUrl },
-    llmProcessor
-  );
-  if (llmResult.processed) {
-    mcpLog("debug", `[${index}/${total}] LLM extraction complete`, "reddit");
-    const header = `## LLM Analysis: ${result.post.title}
-
-**r/${result.post.subreddit}** \u2022 u/${result.post.author} \u2022 \u2B06\uFE0F ${result.post.score} \u2022 \u{1F4AC} ${result.post.commentCount} comments
-\u{1F517} ${result.post.url}
-
-`;
-    return { content: header + llmResult.content, llmFailed: false };
-  }
-  mcpLog("warning", `[${index}/${total}] LLM extraction failed: ${llmResult.error || "unknown"}`, "reddit");
-  return { content: postContent, llmFailed: true };
-}
-async function fetchAndProcessPosts(results, urls, fetchComments, what_to_extract) {
-  const llmProcessor = createLLMProcessor();
-  const tokensPerUrl = Math.floor(TOKEN_BUDGETS.RESEARCH / urls.length);
-  const enhancedInstruction = enhanceExtractionInstruction(what_to_extract);
+async function fetchAndProcessPosts(results) {
   let failed = 0;
   const failedContents = [];
-  const successEntries = [];
-  const skippedUrls = [];
+  const successContents = [];
+  let successful = 0;
   let totalWordsUsed = 0;
+  const skippedUrls = [];
   for (const [url, result] of results) {
     if (result instanceof Error) {
       failed++;
@@ -207,36 +176,13 @@ _${result.message}_`);
       skippedUrls.push(url);
       continue;
     }
-    const formatted = formatPost(result, fetchComments, REDDIT.MAX_WORDS_PER_POST);
+    const formatted = formatPost(result, true, REDDIT.MAX_WORDS_PER_POST);
     totalWordsUsed += formatted.wordsUsed;
-    successEntries.push({ url, result, content: formatted.md, wordsUsed: formatted.wordsUsed });
-  }
-  let llmErrors = 0;
-  let processedEntries;
-  if (llmProcessor && successEntries.length > 0) {
-    const llmResults = await pMap(successEntries, async (entry, index) => {
-      const llmOut = await applyLlmToPost(
-        entry.content,
-        entry.result,
-        entry.url,
-        llmProcessor,
-        enhancedInstruction,
-        tokensPerUrl,
-        index + 1,
-        successEntries.length
-      );
-      if (llmOut.llmFailed) llmErrors++;
-      return { ...entry, content: llmOut.content };
-    }, CONCURRENCY.LLM_EXTRACTION);
-    processedEntries = llmResults;
-  } else {
-    if (!llmProcessor) {
-      mcpLog("warning", "LLM unavailable (LLM_EXTRACTION_API_KEY not set). Returning raw content.", "reddit");
-    }
-    processedEntries = successEntries;
+    successContents.push(formatted.md);
+    successful++;
   }
-  const contents = [...failedContents, ...processedEntries.map((e) => e.content)];
-  return { successful: successEntries.length, failed, llmErrors, llmAvailable: llmProcessor !== null, contents, totalWordsUsed, skippedUrls };
+  const contents = [...failedContents, ...successContents];
+  return { successful, failed, llmErrors: 0, llmAvailable: false, contents, totalWordsUsed, skippedUrls };
 }
 function buildRedditStatusExtras(rateLimitHits, llmAvailable, llmErrors) {
   const extras = [];
@@ -296,56 +242,40 @@ function formatGetRedditPostsError(error) {
     ]
   });
 }
-async function handleGetRedditPosts(urls, clientId, clientSecret, options, reporter = NOOP_REPORTER) {
+async function handleGetRedditPosts(urls, clientId, clientSecret, reporter = NOOP_REPORTER) {
+  const startTime = Date.now();
   try {
-    const { fetchComments = true, what_to_extract } = options;
     const validationError = validatePostCount(urls.length);
     if (validationError) return toolFailure(validationError);
     const totalBatches = Math.ceil(urls.length / REDDIT.BATCH_SIZE);
-    await reporter.log("info", `Fetching ${urls.length} Reddit post(s) across ${totalBatches} batch(es)`);
-    await reporter.progress(20, 100, "Fetching Reddit post content");
+    await reporter.log("info", `Fetching ${urls.length} Reddit post(s)`);
+    await reporter.progress(20, 100, "Fetching Reddit posts");
     const client = new RedditClient(clientId, clientSecret);
-    const batchResult = await client.batchGetPosts(urls, fetchComments);
-    await reporter.log(
-      "info",
-      `Fetched Reddit batch results with ${batchResult.rateLimitHits} rate-limit retry/retries`
-    );
-    await reporter.progress(55, 100, "Processing Reddit posts and LLM extraction");
-    const processResult = await fetchAndProcessPosts(
-      batchResult.results,
-      urls,
-      fetchComments,
-      what_to_extract
-    );
-    await reporter.log(
-      "info",
-      `Processed ${processResult.successful} successful post(s) with ${processResult.failed} failure(s), ${processResult.totalWordsUsed.toLocaleString()} words`
-    );
-    await reporter.progress(85, 100, "Formatting Reddit output");
-    const tokensPerUrl = Math.floor(TOKEN_BUDGETS.RESEARCH / urls.length);
+    const batchResult = await client.batchGetPosts(urls, true);
+    await reporter.progress(55, 100, "Formatting posts and comments");
+    const processResult = await fetchAndProcessPosts(batchResult.results);
+    await reporter.progress(85, 100, "Building output");
     const extraStatus = buildRedditStatusExtras(
       batchResult.rateLimitHits,
-      processResult.llmAvailable,
-      processResult.llmErrors
+      false,
+      0
     );
     const content = formatRedditOutput(
       urls,
       processResult,
-      fetchComments,
+      true,
       totalBatches,
-      tokensPerUrl,
+      0,
       extraStatus
     );
+    const executionTime = Date.now() - startTime;
     return toolSuccess(content, {
       content,
       metadata: {
-        total_urls: urls.length,
+        total_items: urls.length,
         successful: processResult.successful,
         failed: processResult.failed,
-        fetch_comments: fetchComments,
-        total_words_used: processResult.totalWordsUsed,
-        llm_failures: processResult.llmErrors,
-        total_batches: totalBatches,
+        execution_time_ms: executionTime,
         rate_limit_hits: batchResult.rateLimitHits
       }
     });
@@ -358,7 +288,7 @@ function registerSearchRedditTool(server) {
     {
       name: "search-reddit",
       title: "Search Reddit",
-      description: 'Search for Reddit posts by appending "site:reddit.com" to 1-50 queries via Google. Returns a flat list of unique Reddit URLs. No ranking, no LLM processing \u2014 just URL discovery. Pipe results into get-reddit-post to fetch and analyze the actual content.',
+      description: "Search Google for Reddit posts matching up to 100 queries. Returns a flat list of unique Reddit URLs ready to pipe into get-reddit-post.",
       schema: searchRedditParamsSchema,
       outputSchema: searchRedditOutputSchema,
       annotations: {
@@ -385,7 +315,7 @@ function registerGetRedditPostTool(server) {
     {
       name: "get-reddit-post",
       title: "Get Reddit Post",
-      description: 'Fetch 1-50 Reddit posts with full comment trees and run LLM extraction. Provide what_to_extract with specific instructions (e.g., "Extract recommendations | pain points | consensus opinions"). The LLM synthesizes posts and comments into focused insights. Best used after search-reddit.',
+      description: "Fetch up to 100 Reddit posts with full threaded comment trees. Returns the raw post content and all comments with author, score, and OP markers.",
       schema: getRedditPostParamsSchema,
       outputSchema: getRedditPostOutputSchema,
       annotations: {
@@ -395,7 +325,7 @@ function registerGetRedditPostTool(server) {
         openWorldHint: true
       }
     },
-    async ({ urls, fetch_comments, what_to_extract }, ctx) => {
+    async ({ urls }, ctx) => {
       if (!getCapabilities().reddit) {
         return toToolResponse(toolFailure(getMissingEnvMessage("reddit")));
       }
@@ -405,7 +335,6 @@ function registerGetRedditPostTool(server) {
         urls,
         env.REDDIT_CLIENT_ID,
         env.REDDIT_CLIENT_SECRET,
-        { fetchComments: fetch_comments, what_to_extract },
         reporter
       );
       await reporter.progress(100, 100, result.isError ? "Reddit fetch failed" : "Reddit fetch complete");

package/dist/src/tools/reddit.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../src/tools/reddit.ts"],
-  "sourcesContent": ["/**\n * Reddit Tools - Search and Fetch\n * NEVER throws - always returns structured response for graceful degradation\n */\n\nimport type { MCPServer } from 'mcp-use/server';\n\nimport {\n  searchRedditParamsSchema,\n  searchRedditOutputSchema,\n  getRedditPostParamsSchema,\n  getRedditPostOutputSchema,\n  type SearchRedditOutput,\n  type GetRedditPostOutput,\n} from '../schemas/reddit.js';\nimport { SearchClient } from '../clients/search.js';\nimport { RedditClient, type PostResult, type Comment } from '../clients/reddit.js';\nimport { REDDIT, CONCURRENCY, getCapabilities, getMissingEnvMessage, parseEnv } from '../config/index.js';\nimport { classifyError } from '../utils/errors.js';\nimport { createLLMProcessor, processContentWithLLM } from '../services/llm-processor.js';\nimport { pMap } from '../utils/concurrency.js';\nimport {\n  mcpLog,\n  formatSuccess,\n  formatError,\n  formatBatchHeader,\n  TOKEN_BUDGETS,\n} from './utils.js';\nimport {\n  createToolReporter,\n  NOOP_REPORTER,\n  toolFailure,\n  toolSuccess,\n  toToolResponse,\n  type ToolExecutionResult,\n  type ToolReporter,\n} from './mcp-helpers.js';\n\n// ============================================================================\n// Formatters\n// ============================================================================\n\nfunction countWords(text: string): number {\n  const plain = text.replace(/[*_~`#>|[\\]()!-]/g, '');\n  return plain.split(/\\s+/).filter(w => w.length > 0).length;\n}\n\ninterface FormattedCommentsResult {\n  md: string;\n  wordsUsed: number;\n  shown: number;\n  truncated: number;\n}\n\nfunction formatComments(comments: Comment[], maxWords: number): FormattedCommentsResult {\n  let md = '';\n  let wordsUsed = 0;\n  let shown = 0;\n\n  for (const c of comments) {\n    const indent = '  '.repeat(c.depth);\n    const op = c.isOP ? ' **[OP]**' : '';\n    const score = c.score >= 0 ? `+${c.score}` : `${c.score}`;\n    const authorLine = `${indent}- **u/${c.author}**${op} _(${score})_\\n`;\n    const bodyLines = c.body.split('\\n').map(line => `${indent}  ${line}`).join('\\n');\n    const commentMd = `${authorLine}${bodyLines}\\n\\n`;\n    const commentWords = countWords(commentMd);\n\n    if (wordsUsed + commentWords > maxWords && shown > 0) break;\n\n    md += commentMd;\n    wordsUsed += commentWords;\n    shown++;\n  }\n\n  return { md, wordsUsed, shown, truncated: comments.length - shown };\n}\n\ninterface FormattedPostResult {\n  md: string;\n  wordsUsed: number;\n  commentsShown: number;\n  commentsTruncated: number;\n}\n\nfunction formatPost(result: PostResult, fetchComments: boolean, maxWords: number): FormattedPostResult {\n  const { post, comments } = result;\n  let md = `## ${post.title}\\n\\n`;\n  md += `**r/${post.subreddit}** \u2022 u/${post.author} \u2022 \u2B06\uFE0F ${post.score} \u2022 \uD83D\uDCAC ${post.commentCount} comments\\n`;\n  md += `\uD83D\uDD17 ${post.url}\\n\\n`;\n\n  let wordsUsed = countWords(md);\n\n  if (post.body) {\n    const bodySection = `### Post Content\\n\\n${post.body}\\n\\n`;\n    wordsUsed += countWords(bodySection);\n    md += bodySection;\n  }\n\n  let commentsShown = 0;\n  let commentsTruncated = 0;\n\n  if (fetchComments && comments.length > 0) {\n    const remainingWords = Math.max(0, maxWords - wordsUsed);\n    const commentsResult = formatComments(comments, remainingWords);\n    commentsShown = commentsResult.shown;\n    commentsTruncated = commentsResult.truncated;\n\n    md += `### Top Comments (${commentsResult.shown}/${post.commentCount} shown, ${commentsResult.wordsUsed.toLocaleString()} words)\\n\\n`;\n    md += commentsResult.md;\n    wordsUsed += commentsResult.wordsUsed;\n\n    if (commentsResult.truncated > 0) {\n      md += `\\n_${commentsResult.truncated} more comments not shown (word budget reached). Use use_llm=true for AI-synthesized summary._\\n\\n`;\n    }\n  } else if (!fetchComments) {\n    md += `_Comments not fetched (fetch_comments=false)_\\n\\n`;\n  }\n\n  return { md, wordsUsed, commentsShown, commentsTruncated };\n}\n\n// ============================================================================\n// Search Reddit Handler (simplified \u2014 returns flat URL list)\n// ============================================================================\n\nexport async function handleSearchReddit(\n  queries: string[],\n  apiKey: string,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<SearchRedditOutput>> {\n  try {\n    const limited = queries.slice(0, 50);\n    const client = new SearchClient(apiKey);\n    await reporter.log('info', `Searching Reddit with ${limited.length} queries`);\n    await reporter.progress(15, 100, 'Searching Reddit');\n    const results = await client.searchRedditMultiple(limited);\n\n    // Collect all unique URLs\n    const allUrls = new Set<string>();\n    for (const resultSet of results.values()) {\n      for (const result of resultSet) {\n        if (result.url) allUrls.add(result.url);\n      }\n    }\n\n    if (allUrls.size === 0) {\n      return toolFailure(formatError({\n        code: 'NO_RESULTS',\n        message: `No Reddit URLs found for any of the ${limited.length} queries`,\n        toolName: 'search-reddit',\n        howToFix: ['Try broader or simpler search terms', 'Check spelling'],\n        alternatives: ['web-search(keywords=[\"topic reddit discussion\"], objective=\"...\") \u2014 broader Google search'],\n      }));\n    }\n\n    const urlList = [...allUrls];\n    const content = urlList.join('\\n');\n\n    await reporter.log('info', `Found ${urlList.length} unique Reddit URLs across ${limited.length} queries`);\n    await reporter.progress(100, 100, 'Reddit search complete');\n\n    return toolSuccess(content, {\n      content,\n      metadata: {\n        query_count: limited.length,\n        total_urls: urlList.length,\n      },\n    });\n  } catch (error) {\n    const structuredError = classifyError(error);\n    return toolFailure(formatError({\n      code: structuredError.code,\n      message: structuredError.message,\n      retryable: structuredError.retryable,\n      toolName: 'search-reddit',\n      howToFix: ['Verify SERPER_API_KEY is set correctly'],\n    }));\n  }\n}\n\n// ============================================================================\n// Get Reddit Posts Handler\n// ============================================================================\n\ninterface GetRedditPostsOptions {\n  fetchComments?: boolean;\n  what_to_extract: string;\n}\n\nfunction enhanceExtractionInstruction(instruction: string): string {\n  return `${instruction}\\n\\n${REDDIT.EXTRACTION_SUFFIX}`;\n}\n\n// --- Internal types ---\n\ninterface PostProcessResult {\n  successful: number;\n  failed: number;\n  llmErrors: number;\n  llmAvailable: boolean;\n  contents: string[];\n  totalWordsUsed: number;\n  skippedUrls: string[];\n}\n\n// --- Helpers ---\n\nfunction validatePostCount(urlCount: number): string | null {\n  if (urlCount < REDDIT.MIN_POSTS) {\n    return formatError({\n      code: 'MIN_POSTS',\n      message: `Minimum ${REDDIT.MIN_POSTS} Reddit posts required. Received: ${urlCount}`,\n      toolName: 'get-reddit-post',\n      howToFix: [`Add at least ${REDDIT.MIN_POSTS - urlCount} more Reddit URL(s)`],\n      alternatives: [\n        `search-reddit(queries=[\"topic discussion\", \"topic recommendations\", \"topic experiences\"]) \u2014 find more Reddit posts first, then call get-reddit-post with ${REDDIT.MIN_POSTS}+ URLs`,\n      ],\n    });\n  }\n  if (urlCount > REDDIT.MAX_POSTS) {\n    return formatError({\n      code: 'MAX_POSTS',\n      message: `Maximum ${REDDIT.MAX_POSTS} Reddit posts allowed. Received: ${urlCount}`,\n      toolName: 'get-reddit-post',\n      howToFix: [`Remove ${urlCount - REDDIT.MAX_POSTS} URL(s) and retry`],\n    });\n  }\n  return null;\n}\n\nasync function applyLlmToPost(\n  postContent: string,\n  result: PostResult,\n  url: string,\n  llmProcessor: NonNullable<ReturnType<typeof createLLMProcessor>>,\n  enhancedInstruction: string | undefined,\n  tokensPerUrl: number,\n  index: number,\n  total: number,\n): Promise<{ content: string; llmFailed: boolean }> {\n  mcpLog('info', `[${index}/${total}] Applying LLM extraction to ${url}`, 'reddit');\n\n  const llmResult = await processContentWithLLM(\n    postContent,\n    { use_llm: true, what_to_extract: enhancedInstruction, max_tokens: tokensPerUrl },\n    llmProcessor,\n  );\n\n  if (llmResult.processed) {\n    mcpLog('debug', `[${index}/${total}] LLM extraction complete`, 'reddit');\n    const header = `## LLM Analysis: ${result.post.title}\\n\\n**r/${result.post.subreddit}** \u2022 u/${result.post.author} \u2022 \u2B06\uFE0F ${result.post.score} \u2022 \uD83D\uDCAC ${result.post.commentCount} comments\\n\uD83D\uDD17 ${result.post.url}\\n\\n`;\n    return { content: header + llmResult.content, llmFailed: false };\n  }\n\n  mcpLog('warning', `[${index}/${total}] LLM extraction failed: ${llmResult.error || 'unknown'}`, 'reddit');\n  return { content: postContent, llmFailed: true };\n}\n\nasync function fetchAndProcessPosts(\n  results: Map<string, PostResult | Error>,\n  urls: string[],\n  fetchComments: boolean,\n  what_to_extract: string,\n): Promise<PostProcessResult> {\n  const llmProcessor = createLLMProcessor();\n  const tokensPerUrl = Math.floor(TOKEN_BUDGETS.RESEARCH / urls.length);\n  const enhancedInstruction = enhanceExtractionInstruction(what_to_extract);\n\n  let failed = 0;\n  const failedContents: string[] = [];\n  const successEntries: { url: string; result: PostResult; content: string; wordsUsed: number }[] = [];\n  const skippedUrls: string[] = [];\n  let totalWordsUsed = 0;\n\n  for (const [url, result] of results) {\n    if (result instanceof Error) {\n      failed++;\n      failedContents.push(`## \u274C Failed: ${url}\\n\\n_${result.message}_`);\n      continue;\n    }\n\n    if (totalWordsUsed >= REDDIT.MAX_WORDS_TOTAL) {\n      skippedUrls.push(url);\n      continue;\n    }\n\n    const formatted = formatPost(result, fetchComments, REDDIT.MAX_WORDS_PER_POST);\n    totalWordsUsed += formatted.wordsUsed;\n    successEntries.push({ url, result, content: formatted.md, wordsUsed: formatted.wordsUsed });\n  }\n\n  let llmErrors = 0;\n  let processedEntries: typeof successEntries;\n\n  // Always run LLM when available\n  if (llmProcessor && successEntries.length > 0) {\n    const llmResults = await pMap(successEntries, async (entry, index) => {\n      const llmOut = await applyLlmToPost(\n        entry.content, entry.result, entry.url, llmProcessor, enhancedInstruction,\n        tokensPerUrl, index + 1, successEntries.length,\n      );\n      if (llmOut.llmFailed) llmErrors++;\n      return { ...entry, content: llmOut.content };\n    }, CONCURRENCY.LLM_EXTRACTION);\n    processedEntries = llmResults;\n  } else {\n    if (!llmProcessor) {\n      mcpLog('warning', 'LLM unavailable (LLM_EXTRACTION_API_KEY not set). Returning raw content.', 'reddit');\n    }\n    processedEntries = successEntries;\n  }\n\n  const contents = [...failedContents, ...processedEntries.map(e => e.content)];\n\n  return { successful: successEntries.length, failed, llmErrors, llmAvailable: llmProcessor !== null, contents, totalWordsUsed, skippedUrls };\n}\n\nfunction buildRedditStatusExtras(\n  rateLimitHits: number,\n  llmAvailable: boolean,\n  llmErrors: number,\n): string {\n  const extras: string[] = [];\n  if (rateLimitHits > 0) extras.push(`\u26A0\uFE0F ${rateLimitHits} rate limit retries`);\n  if (!llmAvailable) {\n    extras.push('\u26A0\uFE0F LLM unavailable (LLM_EXTRACTION_API_KEY not set) \u2014 raw content returned');\n  } else if (llmErrors > 0) {\n    extras.push(`\u26A0\uFE0F ${llmErrors} LLM extraction failures`);\n  }\n  return extras.length > 0 ? `\\n${extras.join(' | ')}` : '';\n}\n\nfunction formatRedditOutput(\n  urls: string[],\n  processResult: PostProcessResult,\n  fetchComments: boolean,\n  totalBatches: number,\n  tokensPerUrl: number,\n  extraStatus: string,\n): string {\n  const batchHeader = formatBatchHeader({\n    title: `Reddit Posts`,\n    totalItems: urls.length,\n    successful: processResult.successful,\n    failed: processResult.failed,\n    ...(fetchComments ? { extras: { 'Words used': processResult.totalWordsUsed.toLocaleString() } } : {}),\n    tokensPerItem: tokensPerUrl,\n    batches: totalBatches,\n  });\n\n  let data = processResult.contents.join('\\n\\n---\\n\\n');\n\n  if (processResult.skippedUrls.length > 0) {\n    data += '\\n\\n---\\n\\n';\n    data += `**Word limit reached (${REDDIT.MAX_WORDS_TOTAL.toLocaleString()} words).** The following posts were not included:\\n`;\n    for (const url of processResult.skippedUrls) {\n      data += `- ${url}\\n`;\n    }\n    data += `\\nCall get-reddit-post again with just these URLs.`;\n  }\n\n  return formatSuccess({\n    title: `Reddit Posts Fetched (${processResult.successful}/${urls.length})`,\n    summary: batchHeader + extraStatus,\n    data,\n    nextSteps: [\n      processResult.successful > 0 ? 'web-search to verify claims from Reddit discussions' : null,\n      processResult.successful > 0 ? 'scrape-links on URLs referenced in comments' : null,\n      processResult.failed > 0 ? 'Retry failed URLs individually' : null,\n    ].filter(Boolean) as string[],\n  });\n}\n\nfunction formatGetRedditPostsError(error: unknown): string {\n  const structuredError = classifyError(error);\n  return formatError({\n    code: structuredError.code,\n    message: structuredError.message,\n    retryable: structuredError.retryable,\n    toolName: 'get-reddit-post',\n    howToFix: ['Verify REDDIT_CLIENT_ID and REDDIT_CLIENT_SECRET are set'],\n    alternatives: [\n      'web-search(keywords=[\"topic reddit discussion\", \"topic reddit recommendations\"]) \u2014 search for Reddit content via web search instead',\n      'scrape-links(urls=[...the Reddit URLs...], use_llm=true, what_to_extract=\"Extract post content | top comments | recommendations\") \u2014 scrape Reddit pages directly as a fallback',\n    ],\n  });\n}\n\nexport async function handleGetRedditPosts(\n  urls: string[],\n  clientId: string,\n  clientSecret: string,\n  options: GetRedditPostsOptions,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<GetRedditPostOutput>> {\n  try {\n    const { fetchComments = true, what_to_extract } = options;\n\n    const validationError = validatePostCount(urls.length);\n    if (validationError) return toolFailure(validationError);\n\n    const totalBatches = Math.ceil(urls.length / REDDIT.BATCH_SIZE);\n\n    await reporter.log('info', `Fetching ${urls.length} Reddit post(s) across ${totalBatches} batch(es)`);\n    await reporter.progress(20, 100, 'Fetching Reddit post content');\n    const client = new RedditClient(clientId, clientSecret);\n    const batchResult = await client.batchGetPosts(urls, fetchComments);\n    await reporter.log(\n      'info',\n      `Fetched Reddit batch results with ${batchResult.rateLimitHits} rate-limit retry/retries`,\n    );\n    await reporter.progress(55, 100, 'Processing Reddit posts and LLM extraction');\n\n    const processResult = await fetchAndProcessPosts(\n      batchResult.results, urls, fetchComments, what_to_extract,\n    );\n    await reporter.log(\n      'info',\n      `Processed ${processResult.successful} successful post(s) with ${processResult.failed} failure(s), ${processResult.totalWordsUsed.toLocaleString()} words`,\n    );\n    await reporter.progress(85, 100, 'Formatting Reddit output');\n\n    const tokensPerUrl = Math.floor(TOKEN_BUDGETS.RESEARCH / urls.length);\n    const extraStatus = buildRedditStatusExtras(\n      batchResult.rateLimitHits, processResult.llmAvailable, processResult.llmErrors,\n    );\n    const content = formatRedditOutput(\n      urls,\n      processResult,\n      fetchComments,\n      totalBatches,\n      tokensPerUrl,\n      extraStatus,\n    );\n\n    return toolSuccess(content, {\n      content,\n      metadata: {\n        total_urls: urls.length,\n        successful: processResult.successful,\n        failed: processResult.failed,\n        fetch_comments: fetchComments,\n        total_words_used: processResult.totalWordsUsed,\n        llm_failures: processResult.llmErrors,\n        total_batches: totalBatches,\n        rate_limit_hits: batchResult.rateLimitHits,\n      },\n    });\n  } catch (error) {\n    return toolFailure(formatGetRedditPostsError(error));\n  }\n}\n\nexport function registerSearchRedditTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'search-reddit',\n      title: 'Search Reddit',\n      description:\n        'Search for Reddit posts by appending \"site:reddit.com\" to 1-50 queries via Google. Returns a flat list of unique Reddit URLs. No ranking, no LLM processing \u2014 just URL discovery. Pipe results into get-reddit-post to fetch and analyze the actual content.',\n      schema: searchRedditParamsSchema,\n      outputSchema: searchRedditOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async ({ queries }, ctx) => {\n      if (!getCapabilities().search) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('search')));\n      }\n\n      const env = parseEnv();\n      const reporter = createToolReporter(ctx, 'search-reddit');\n      const result = await handleSearchReddit(queries, env.SEARCH_API_KEY!, reporter);\n\n      await reporter.progress(100, 100, result.isError ? 'Reddit search failed' : 'Reddit search complete');\n      return toToolResponse(result);\n    },\n  );\n}\n\nexport function registerGetRedditPostTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'get-reddit-post',\n      title: 'Get Reddit Post',\n      description:\n        'Fetch 1-50 Reddit posts with full comment trees and run LLM extraction. Provide what_to_extract with specific instructions (e.g., \"Extract recommendations | pain points | consensus opinions\"). The LLM synthesizes posts and comments into focused insights. Best used after search-reddit.',\n      schema: getRedditPostParamsSchema,\n      outputSchema: getRedditPostOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async ({ urls, fetch_comments, what_to_extract }, ctx) => {\n      if (!getCapabilities().reddit) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('reddit')));\n      }\n\n      const env = parseEnv();\n      const reporter = createToolReporter(ctx, 'get-reddit-post');\n      const result = await handleGetRedditPosts(\n        urls,\n        env.REDDIT_CLIENT_ID!,\n        env.REDDIT_CLIENT_SECRET!,\n        { fetchComments: fetch_comments, what_to_extract },\n        reporter,\n      );\n\n      await reporter.progress(100, 100, result.isError ? 'Reddit fetch failed' : 'Reddit fetch complete');\n      return toToolResponse(result);\n    },\n  );\n}\n"],
-  "mappings": "AAOA;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AACP,SAAS,oBAAoB;AAC7B,SAAS,oBAAmD;AAC5D,SAAS,QAAQ,aAAa,iBAAiB,sBAAsB,gBAAgB;AACrF,SAAS,qBAAqB;AAC9B,SAAS,oBAAoB,6BAA6B;AAC1D,SAAS,YAAY;AACrB;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OACK;AACP;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AAMP,SAAS,WAAW,MAAsB;AACxC,QAAM,QAAQ,KAAK,QAAQ,qBAAqB,EAAE;AAClD,SAAO,MAAM,MAAM,KAAK,EAAE,OAAO,OAAK,EAAE,SAAS,CAAC,EAAE;AACtD;AASA,SAAS,eAAe,UAAqB,UAA2C;AACtF,MAAI,KAAK;AACT,MAAI,YAAY;AAChB,MAAI,QAAQ;AAEZ,aAAW,KAAK,UAAU;AACxB,UAAM,SAAS,KAAK,OAAO,EAAE,KAAK;AAClC,UAAM,KAAK,EAAE,OAAO,cAAc;AAClC,UAAM,QAAQ,EAAE,SAAS,IAAI,IAAI,EAAE,KAAK,KAAK,GAAG,EAAE,KAAK;AACvD,UAAM,aAAa,GAAG,MAAM,SAAS,EAAE,MAAM,KAAK,EAAE,MAAM,KAAK;AAAA;AAC/D,UAAM,YAAY,EAAE,KAAK,MAAM,IAAI,EAAE,IAAI,UAAQ,GAAG,MAAM,KAAK,IAAI,EAAE,EAAE,KAAK,IAAI;AAChF,UAAM,YAAY,GAAG,UAAU,GAAG,SAAS;AAAA;AAAA;AAC3C,UAAM,eAAe,WAAW,SAAS;AAEzC,QAAI,YAAY,eAAe,YAAY,QAAQ,EAAG;AAEtD,UAAM;AACN,iBAAa;AACb;AAAA,EACF;AAEA,SAAO,EAAE,IAAI,WAAW,OAAO,WAAW,SAAS,SAAS,MAAM;AACpE;AASA,SAAS,WAAW,QAAoB,eAAwB,UAAuC;AACrG,QAAM,EAAE,MAAM,SAAS,IAAI;AAC3B,MAAI,KAAK,MAAM,KAAK,KAAK;AAAA;AAAA;AACzB,QAAM,OAAO,KAAK,SAAS,eAAU,KAAK,MAAM,wBAAS,KAAK,KAAK,qBAAS,KAAK,YAAY;AAAA;AAC7F,QAAM,aAAM,KAAK,GAAG;AAAA;AAAA;AAEpB,MAAI,YAAY,WAAW,EAAE;AAE7B,MAAI,KAAK,MAAM;AACb,UAAM,cAAc;AAAA;AAAA,EAAuB,KAAK,IAAI;AAAA;AAAA;AACpD,iBAAa,WAAW,WAAW;AACnC,UAAM;AAAA,EACR;AAEA,MAAI,gBAAgB;AACpB,MAAI,oBAAoB;AAExB,MAAI,iBAAiB,SAAS,SAAS,GAAG;AACxC,UAAM,iBAAiB,KAAK,IAAI,GAAG,WAAW,SAAS;AACvD,UAAM,iBAAiB,eAAe,UAAU,cAAc;AAC9D,oBAAgB,eAAe;AAC/B,wBAAoB,eAAe;AAEnC,UAAM,qBAAqB,eAAe,KAAK,IAAI,KAAK,YAAY,WAAW,eAAe,UAAU,eAAe,CAAC;AAAA;AAAA;AACxH,UAAM,eAAe;AACrB,iBAAa,eAAe;AAE5B,QAAI,eAAe,YAAY,GAAG;AAChC,YAAM;AAAA,GAAM,eAAe,SAAS;AAAA;AAAA;AAAA,IACtC;AAAA,EACF,WAAW,CAAC,eAAe;AACzB,UAAM;AAAA;AAAA;AAAA,EACR;AAEA,SAAO,EAAE,IAAI,WAAW,eAAe,kBAAkB;AAC3D;AAMA,eAAsB,mBACpB,SACA,QACA,WAAyB,eACyB;AAClD,MAAI;AACF,UAAM,UAAU,QAAQ,MAAM,GAAG,EAAE;AACnC,UAAM,SAAS,IAAI,aAAa,MAAM;AACtC,UAAM,SAAS,IAAI,QAAQ,yBAAyB,QAAQ,MAAM,UAAU;AAC5E,UAAM,SAAS,SAAS,IAAI,KAAK,kBAAkB;AACnD,UAAM,UAAU,MAAM,OAAO,qBAAqB,OAAO;AAGzD,UAAM,UAAU,oBAAI,IAAY;AAChC,eAAW,aAAa,QAAQ,OAAO,GAAG;AACxC,iBAAW,UAAU,WAAW;AAC9B,YAAI,OAAO,IAAK,SAAQ,IAAI,OAAO,GAAG;AAAA,MACxC;AAAA,IACF;AAEA,QAAI,QAAQ,SAAS,GAAG;AACtB,aAAO,YAAY,YAAY;AAAA,QAC7B,MAAM;AAAA,QACN,SAAS,uCAAuC,QAAQ,MAAM;AAAA,QAC9D,UAAU;AAAA,QACV,UAAU,CAAC,uCAAuC,gBAAgB;AAAA,QAClE,cAAc,CAAC,gGAA2F;AAAA,MAC5G,CAAC,CAAC;AAAA,IACJ;AAEA,UAAM,UAAU,CAAC,GAAG,OAAO;AAC3B,UAAM,UAAU,QAAQ,KAAK,IAAI;AAEjC,UAAM,SAAS,IAAI,QAAQ,SAAS,QAAQ,MAAM,8BAA8B,QAAQ,MAAM,UAAU;AACxG,UAAM,SAAS,SAAS,KAAK,KAAK,wBAAwB;AAE1D,WAAO,YAAY,SAAS;AAAA,MAC1B;AAAA,MACA,UAAU;AAAA,QACR,aAAa,QAAQ;AAAA,QACrB,YAAY,QAAQ;AAAA,MACtB;AAAA,IACF,CAAC;AAAA,EACH,SAAS,OAAO;AACd,UAAM,kBAAkB,cAAc,KAAK;AAC3C,WAAO,YAAY,YAAY;AAAA,MAC7B,MAAM,gBAAgB;AAAA,MACtB,SAAS,gBAAgB;AAAA,MACzB,WAAW,gBAAgB;AAAA,MAC3B,UAAU;AAAA,MACV,UAAU,CAAC,wCAAwC;AAAA,IACrD,CAAC,CAAC;AAAA,EACJ;AACF;AAWA,SAAS,6BAA6B,aAA6B;AACjE,SAAO,GAAG,WAAW;AAAA;AAAA,EAAO,OAAO,iBAAiB;AACtD;AAgBA,SAAS,kBAAkB,UAAiC;AAC1D,MAAI,WAAW,OAAO,WAAW;AAC/B,WAAO,YAAY;AAAA,MACjB,MAAM;AAAA,MACN,SAAS,WAAW,OAAO,SAAS,qCAAqC,QAAQ;AAAA,MACjF,UAAU;AAAA,MACV,UAAU,CAAC,gBAAgB,OAAO,YAAY,QAAQ,qBAAqB;AAAA,MAC3E,cAAc;AAAA,QACZ,iKAA4J,OAAO,SAAS;AAAA,MAC9K;AAAA,IACF,CAAC;AAAA,EACH;AACA,MAAI,WAAW,OAAO,WAAW;AAC/B,WAAO,YAAY;AAAA,MACjB,MAAM;AAAA,MACN,SAAS,WAAW,OAAO,SAAS,oCAAoC,QAAQ;AAAA,MAChF,UAAU;AAAA,MACV,UAAU,CAAC,UAAU,WAAW,OAAO,SAAS,mBAAmB;AAAA,IACrE,CAAC;AAAA,EACH;AACA,SAAO;AACT;AAEA,eAAe,eACb,aACA,QACA,KACA,cACA,qBACA,cACA,OACA,OACkD;AAClD,SAAO,QAAQ,IAAI,KAAK,IAAI,KAAK,gCAAgC,GAAG,IAAI,QAAQ;AAEhF,QAAM,YAAY,MAAM;AAAA,IACtB;AAAA,IACA,EAAE,SAAS,MAAM,iBAAiB,qBAAqB,YAAY,aAAa;AAAA,IAChF;AAAA,EACF;AAEA,MAAI,UAAU,WAAW;AACvB,WAAO,SAAS,IAAI,KAAK,IAAI,KAAK,6BAA6B,QAAQ;AACvE,UAAM,SAAS,oBAAoB,OAAO,KAAK,KAAK;AAAA;AAAA,MAAW,OAAO,KAAK,SAAS,eAAU,OAAO,KAAK,MAAM,wBAAS,OAAO,KAAK,KAAK,qBAAS,OAAO,KAAK,YAAY;AAAA,YAAiB,OAAO,KAAK,GAAG;AAAA;AAAA;AAC3M,WAAO,EAAE,SAAS,SAAS,UAAU,SAAS,WAAW,MAAM;AAAA,EACjE;AAEA,SAAO,WAAW,IAAI,KAAK,IAAI,KAAK,4BAA4B,UAAU,SAAS,SAAS,IAAI,QAAQ;AACxG,SAAO,EAAE,SAAS,aAAa,WAAW,KAAK;AACjD;AAEA,eAAe,qBACb,SACA,MACA,eACA,iBAC4B;AAC5B,QAAM,eAAe,mBAAmB;AACxC,QAAM,eAAe,KAAK,MAAM,cAAc,WAAW,KAAK,MAAM;AACpE,QAAM,sBAAsB,6BAA6B,eAAe;AAExE,MAAI,SAAS;AACb,QAAM,iBAA2B,CAAC;AAClC,QAAM,iBAA4F,CAAC;AACnG,QAAM,cAAwB,CAAC;AAC/B,MAAI,iBAAiB;AAErB,aAAW,CAAC,KAAK,MAAM,KAAK,SAAS;AACnC,QAAI,kBAAkB,OAAO;AAC3B;AACA,qBAAe,KAAK,qBAAgB,GAAG;AAAA;AAAA,GAAQ,OAAO,OAAO,GAAG;AAChE;AAAA,IACF;AAEA,QAAI,kBAAkB,OAAO,iBAAiB;AAC5C,kBAAY,KAAK,GAAG;AACpB;AAAA,IACF;AAEA,UAAM,YAAY,WAAW,QAAQ,eAAe,OAAO,kBAAkB;AAC7E,sBAAkB,UAAU;AAC5B,mBAAe,KAAK,EAAE,KAAK,QAAQ,SAAS,UAAU,IAAI,WAAW,UAAU,UAAU,CAAC;AAAA,EAC5F;AAEA,MAAI,YAAY;AAChB,MAAI;AAGJ,MAAI,gBAAgB,eAAe,SAAS,GAAG;AAC7C,UAAM,aAAa,MAAM,KAAK,gBAAgB,OAAO,OAAO,UAAU;AACpE,YAAM,SAAS,MAAM;AAAA,QACnB,MAAM;AAAA,QAAS,MAAM;AAAA,QAAQ,MAAM;AAAA,QAAK;AAAA,QAAc;AAAA,QACtD;AAAA,QAAc,QAAQ;AAAA,QAAG,eAAe;AAAA,MAC1C;AACA,UAAI,OAAO,UAAW;AACtB,aAAO,EAAE,GAAG,OAAO,SAAS,OAAO,QAAQ;AAAA,IAC7C,GAAG,YAAY,cAAc;AAC7B,uBAAmB;AAAA,EACrB,OAAO;AACL,QAAI,CAAC,cAAc;AACjB,aAAO,WAAW,4EAA4E,QAAQ;AAAA,IACxG;AACA,uBAAmB;AAAA,EACrB;AAEA,QAAM,WAAW,CAAC,GAAG,gBAAgB,GAAG,iBAAiB,IAAI,OAAK,EAAE,OAAO,CAAC;AAE5E,SAAO,EAAE,YAAY,eAAe,QAAQ,QAAQ,WAAW,cAAc,iBAAiB,MAAM,UAAU,gBAAgB,YAAY;AAC5I;AAEA,SAAS,wBACP,eACA,cACA,WACQ;AACR,QAAM,SAAmB,CAAC;AAC1B,MAAI,gBAAgB,EAAG,QAAO,KAAK,gBAAM,aAAa,qBAAqB;AAC3E,MAAI,CAAC,cAAc;AACjB,WAAO,KAAK,2FAA4E;AAAA,EAC1F,WAAW,YAAY,GAAG;AACxB,WAAO,KAAK,gBAAM,SAAS,0BAA0B;AAAA,EACvD;AACA,SAAO,OAAO,SAAS,IAAI;AAAA,EAAK,OAAO,KAAK,KAAK,CAAC,KAAK;AACzD;AAEA,SAAS,mBACP,MACA,eACA,eACA,cACA,cACA,aACQ;AACR,QAAM,cAAc,kBAAkB;AAAA,IACpC,OAAO;AAAA,IACP,YAAY,KAAK;AAAA,IACjB,YAAY,cAAc;AAAA,IAC1B,QAAQ,cAAc;AAAA,IACtB,GAAI,gBAAgB,EAAE,QAAQ,EAAE,cAAc,cAAc,eAAe,eAAe,EAAE,EAAE,IAAI,CAAC;AAAA,IACnG,eAAe;AAAA,IACf,SAAS;AAAA,EACX,CAAC;AAED,MAAI,OAAO,cAAc,SAAS,KAAK,aAAa;AAEpD,MAAI,cAAc,YAAY,SAAS,GAAG;AACxC,YAAQ;AACR,YAAQ,yBAAyB,OAAO,gBAAgB,eAAe,CAAC;AAAA;AACxE,eAAW,OAAO,cAAc,aAAa;AAC3C,cAAQ,KAAK,GAAG;AAAA;AAAA,IAClB;AACA,YAAQ;AAAA;AAAA,EACV;AAEA,SAAO,cAAc;AAAA,IACnB,OAAO,yBAAyB,cAAc,UAAU,IAAI,KAAK,MAAM;AAAA,IACvE,SAAS,cAAc;AAAA,IACvB;AAAA,IACA,WAAW;AAAA,MACT,cAAc,aAAa,IAAI,wDAAwD;AAAA,MACvF,cAAc,aAAa,IAAI,gDAAgD;AAAA,MAC/E,cAAc,SAAS,IAAI,mCAAmC;AAAA,IAChE,EAAE,OAAO,OAAO;AAAA,EAClB,CAAC;AACH;AAEA,SAAS,0BAA0B,OAAwB;AACzD,QAAM,kBAAkB,cAAc,KAAK;AAC3C,SAAO,YAAY;AAAA,IACjB,MAAM,gBAAgB;AAAA,IACtB,SAAS,gBAAgB;AAAA,IACzB,WAAW,gBAAgB;AAAA,IAC3B,UAAU;AAAA,IACV,UAAU,CAAC,0DAA0D;AAAA,IACrE,cAAc;AAAA,MACZ;AAAA,MACA;AAAA,IACF;AAAA,EACF,CAAC;AACH;AAEA,eAAsB,qBACpB,MACA,UACA,cACA,SACA,WAAyB,eAC0B;AACnD,MAAI;AACF,UAAM,EAAE,gBAAgB,MAAM,gBAAgB,IAAI;AAElD,UAAM,kBAAkB,kBAAkB,KAAK,MAAM;AACrD,QAAI,gBAAiB,QAAO,YAAY,eAAe;AAEvD,UAAM,eAAe,KAAK,KAAK,KAAK,SAAS,OAAO,UAAU;AAE9D,UAAM,SAAS,IAAI,QAAQ,YAAY,KAAK,MAAM,0BAA0B,YAAY,YAAY;AACpG,UAAM,SAAS,SAAS,IAAI,KAAK,8BAA8B;AAC/D,UAAM,SAAS,IAAI,aAAa,UAAU,YAAY;AACtD,UAAM,cAAc,MAAM,OAAO,cAAc,MAAM,aAAa;AAClE,UAAM,SAAS;AAAA,MACb;AAAA,MACA,qCAAqC,YAAY,aAAa;AAAA,IAChE;AACA,UAAM,SAAS,SAAS,IAAI,KAAK,4CAA4C;AAE7E,UAAM,gBAAgB,MAAM;AAAA,MAC1B,YAAY;AAAA,MAAS;AAAA,MAAM;AAAA,MAAe;AAAA,IAC5C;AACA,UAAM,SAAS;AAAA,MACb;AAAA,MACA,aAAa,cAAc,UAAU,4BAA4B,cAAc,MAAM,gBAAgB,cAAc,eAAe,eAAe,CAAC;AAAA,IACpJ;AACA,UAAM,SAAS,SAAS,IAAI,KAAK,0BAA0B;AAE3D,UAAM,eAAe,KAAK,MAAM,cAAc,WAAW,KAAK,MAAM;AACpE,UAAM,cAAc;AAAA,MAClB,YAAY;AAAA,MAAe,cAAc;AAAA,MAAc,cAAc;AAAA,IACvE;AACA,UAAM,UAAU;AAAA,MACd;AAAA,MACA;AAAA,MACA;AAAA,MACA;AAAA,MACA;AAAA,MACA;AAAA,IACF;AAEA,WAAO,YAAY,SAAS;AAAA,MAC1B;AAAA,MACA,UAAU;AAAA,QACR,YAAY,KAAK;AAAA,QACjB,YAAY,cAAc;AAAA,QAC1B,QAAQ,cAAc;AAAA,QACtB,gBAAgB;AAAA,QAChB,kBAAkB,cAAc;AAAA,QAChC,cAAc,cAAc;AAAA,QAC5B,eAAe;AAAA,QACf,iBAAiB,YAAY;AAAA,MAC/B;AAAA,IACF,CAAC;AAAA,EACH,SAAS,OAAO;AACd,WAAO,YAAY,0BAA0B,KAAK,CAAC;AAAA,EACrD;AACF;AAEO,SAAS,yBAAyB,QAAyB;AAChE,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,EAAE,QAAQ,GAAG,QAAQ;AAC1B,UAAI,CAAC,gBAAgB,EAAE,QAAQ;AAC7B,eAAO,eAAe,YAAY,qBAAqB,QAAQ,CAAC,CAAC;AAAA,MACnE;AAEA,YAAM,MAAM,SAAS;AACrB,YAAM,WAAW,mBAAmB,KAAK,eAAe;AACxD,YAAM,SAAS,MAAM,mBAAmB,SAAS,IAAI,gBAAiB,QAAQ;AAE9E,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,yBAAyB,wBAAwB;AACpG,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;AAEO,SAAS,0BAA0B,QAAyB;AACjE,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,EAAE,MAAM,gBAAgB,gBAAgB,GAAG,QAAQ;AACxD,UAAI,CAAC,gBAAgB,EAAE,QAAQ;AAC7B,eAAO,eAAe,YAAY,qBAAqB,QAAQ,CAAC,CAAC;AAAA,MACnE;AAEA,YAAM,MAAM,SAAS;AACrB,YAAM,WAAW,mBAAmB,KAAK,iBAAiB;AAC1D,YAAM,SAAS,MAAM;AAAA,QACnB;AAAA,QACA,IAAI;AAAA,QACJ,IAAI;AAAA,QACJ,EAAE,eAAe,gBAAgB,gBAAgB;AAAA,QACjD;AAAA,MACF;AAEA,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,wBAAwB,uBAAuB;AAClG,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;",
+  "sourcesContent": ["/**\n * Reddit Tools - Search and Fetch\n * NEVER throws - always returns structured response for graceful degradation\n */\n\nimport type { MCPServer } from 'mcp-use/server';\n\nimport {\n  searchRedditParamsSchema,\n  searchRedditOutputSchema,\n  getRedditPostParamsSchema,\n  getRedditPostOutputSchema,\n  type SearchRedditOutput,\n  type GetRedditPostOutput,\n} from '../schemas/reddit.js';\nimport { SearchClient } from '../clients/search.js';\nimport { RedditClient, type PostResult, type Comment } from '../clients/reddit.js';\nimport { REDDIT, getCapabilities, getMissingEnvMessage, parseEnv } from '../config/index.js';\nimport { classifyError } from '../utils/errors.js';\nimport {\n  mcpLog,\n  formatSuccess,\n  formatError,\n  formatBatchHeader,\n} from './utils.js';\nimport {\n  createToolReporter,\n  NOOP_REPORTER,\n  toolFailure,\n  toolSuccess,\n  toToolResponse,\n  type ToolExecutionResult,\n  type ToolReporter,\n} from './mcp-helpers.js';\n\n// ============================================================================\n// Formatters\n// ============================================================================\n\nfunction countWords(text: string): number {\n  const plain = text.replace(/[*_~`#>|[\\]()!-]/g, '');\n  return plain.split(/\\s+/).filter(w => w.length > 0).length;\n}\n\ninterface FormattedCommentsResult {\n  md: string;\n  wordsUsed: number;\n  shown: number;\n  truncated: number;\n}\n\nfunction formatComments(comments: Comment[], maxWords: number): FormattedCommentsResult {\n  let md = '';\n  let wordsUsed = 0;\n  let shown = 0;\n\n  for (const c of comments) {\n    const indent = '  '.repeat(c.depth);\n    const op = c.isOP ? ' **[OP]**' : '';\n    const score = c.score >= 0 ? `+${c.score}` : `${c.score}`;\n    const authorLine = `${indent}- **u/${c.author}**${op} _(${score})_\\n`;\n    const bodyLines = c.body.split('\\n').map(line => `${indent}  ${line}`).join('\\n');\n    const commentMd = `${authorLine}${bodyLines}\\n\\n`;\n    const commentWords = countWords(commentMd);\n\n    if (wordsUsed + commentWords > maxWords && shown > 0) break;\n\n    md += commentMd;\n    wordsUsed += commentWords;\n    shown++;\n  }\n\n  return { md, wordsUsed, shown, truncated: comments.length - shown };\n}\n\ninterface FormattedPostResult {\n  md: string;\n  wordsUsed: number;\n  commentsShown: number;\n  commentsTruncated: number;\n}\n\nfunction formatPost(result: PostResult, fetchComments: boolean, maxWords: number): FormattedPostResult {\n  const { post, comments } = result;\n  let md = `## ${post.title}\\n\\n`;\n  md += `**r/${post.subreddit}** \u2022 u/${post.author} \u2022 \u2B06\uFE0F ${post.score} \u2022 \uD83D\uDCAC ${post.commentCount} comments\\n`;\n  md += `\uD83D\uDD17 ${post.url}\\n\\n`;\n\n  let wordsUsed = countWords(md);\n\n  if (post.body) {\n    const bodySection = `### Post Content\\n\\n${post.body}\\n\\n`;\n    wordsUsed += countWords(bodySection);\n    md += bodySection;\n  }\n\n  let commentsShown = 0;\n  let commentsTruncated = 0;\n\n  if (fetchComments && comments.length > 0) {\n    const remainingWords = Math.max(0, maxWords - wordsUsed);\n    const commentsResult = formatComments(comments, remainingWords);\n    commentsShown = commentsResult.shown;\n    commentsTruncated = commentsResult.truncated;\n\n    md += `### Top Comments (${commentsResult.shown}/${post.commentCount} shown, ${commentsResult.wordsUsed.toLocaleString()} words)\\n\\n`;\n    md += commentsResult.md;\n    wordsUsed += commentsResult.wordsUsed;\n\n    if (commentsResult.truncated > 0) {\n      md += `\\n_${commentsResult.truncated} more comments not shown (word budget reached)._\\n\\n`;\n    }\n  }\n\n  return { md, wordsUsed, commentsShown, commentsTruncated };\n}\n\n// ============================================================================\n// Search Reddit Handler (simplified \u2014 returns flat URL list)\n// ============================================================================\n\nexport async function handleSearchReddit(\n  queries: string[],\n  apiKey: string,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<SearchRedditOutput>> {\n  try {\n    const startTime = Date.now();\n    const limited = queries.slice(0, 100);\n    const client = new SearchClient(apiKey);\n    await reporter.log('info', `Searching Reddit with ${limited.length} queries`);\n    await reporter.progress(15, 100, 'Searching Reddit');\n    const results = await client.searchRedditMultiple(limited);\n\n    // Collect all unique URLs\n    const allUrls = new Set<string>();\n    for (const resultSet of results.values()) {\n      for (const result of resultSet) {\n        if (result.url) allUrls.add(result.url);\n      }\n    }\n\n    if (allUrls.size === 0) {\n      return toolFailure(formatError({\n        code: 'NO_RESULTS',\n        message: `No Reddit URLs found for any of the ${limited.length} queries`,\n        toolName: 'search-reddit',\n        howToFix: ['Try broader or simpler search terms', 'Check spelling'],\n        alternatives: ['web-search(queries=[\"topic reddit discussion\"], extract=\"...\") \u2014 broader Google search'],\n      }));\n    }\n\n    const urlList = [...allUrls];\n    const content = urlList.join('\\n');\n\n    await reporter.log('info', `Found ${urlList.length} unique Reddit URLs across ${limited.length} queries`);\n    await reporter.progress(100, 100, 'Reddit search complete');\n\n    const executionTime = Date.now() - startTime;\n    return toolSuccess(content, {\n      content,\n      metadata: {\n        total_items: limited.length,\n        successful: urlList.length,\n        failed: 0,\n        execution_time_ms: executionTime,\n      },\n    });\n  } catch (error) {\n    const structuredError = classifyError(error);\n    return toolFailure(formatError({\n      code: structuredError.code,\n      message: structuredError.message,\n      retryable: structuredError.retryable,\n      toolName: 'search-reddit',\n      howToFix: ['Verify SERPER_API_KEY is set correctly'],\n    }));\n  }\n}\n\n// ============================================================================\n// Get Reddit Posts Handler\n// ============================================================================\n\n// get-reddit-post no longer uses LLM \u2014 returns raw posts + comments\n\n// --- Internal types ---\n\ninterface PostProcessResult {\n  successful: number;\n  failed: number;\n  llmErrors: number;\n  llmAvailable: boolean;\n  contents: string[];\n  totalWordsUsed: number;\n  skippedUrls: string[];\n}\n\n// --- Helpers ---\n\nfunction validatePostCount(urlCount: number): string | null {\n  if (urlCount < REDDIT.MIN_POSTS) {\n    return formatError({\n      code: 'MIN_POSTS',\n      message: `Minimum ${REDDIT.MIN_POSTS} Reddit posts required. Received: ${urlCount}`,\n      toolName: 'get-reddit-post',\n      howToFix: [`Add at least ${REDDIT.MIN_POSTS - urlCount} more Reddit URL(s)`],\n      alternatives: [\n        `search-reddit(queries=[\"topic discussion\", \"topic recommendations\", \"topic experiences\"]) \u2014 find more Reddit posts first, then call get-reddit-post with ${REDDIT.MIN_POSTS}+ URLs`,\n      ],\n    });\n  }\n  if (urlCount > REDDIT.MAX_POSTS) {\n    return formatError({\n      code: 'MAX_POSTS',\n      message: `Maximum ${REDDIT.MAX_POSTS} Reddit posts allowed. Received: ${urlCount}`,\n      toolName: 'get-reddit-post',\n      howToFix: [`Remove ${urlCount - REDDIT.MAX_POSTS} URL(s) and retry`],\n    });\n  }\n  return null;\n}\n\nasync function fetchAndProcessPosts(\n  results: Map<string, PostResult | Error>,\n): Promise<PostProcessResult> {\n  let failed = 0;\n  const failedContents: string[] = [];\n  const successContents: string[] = [];\n  let successful = 0;\n  let totalWordsUsed = 0;\n  const skippedUrls: string[] = [];\n\n  for (const [url, result] of results) {\n    if (result instanceof Error) {\n      failed++;\n      failedContents.push(`## \u274C Failed: ${url}\\n\\n_${result.message}_`);\n      continue;\n    }\n\n    if (totalWordsUsed >= REDDIT.MAX_WORDS_TOTAL) {\n      skippedUrls.push(url);\n      continue;\n    }\n\n    const formatted = formatPost(result, true, REDDIT.MAX_WORDS_PER_POST);\n    totalWordsUsed += formatted.wordsUsed;\n    successContents.push(formatted.md);\n    successful++;\n  }\n\n  const contents = [...failedContents, ...successContents];\n  return { successful, failed, llmErrors: 0, llmAvailable: false, contents, totalWordsUsed, skippedUrls };\n}\n\nfunction buildRedditStatusExtras(\n  rateLimitHits: number,\n  llmAvailable: boolean,\n  llmErrors: number,\n): string {\n  const extras: string[] = [];\n  if (rateLimitHits > 0) extras.push(`\u26A0\uFE0F ${rateLimitHits} rate limit retries`);\n  if (!llmAvailable) {\n    extras.push('\u26A0\uFE0F LLM unavailable (LLM_EXTRACTION_API_KEY not set) \u2014 raw content returned');\n  } else if (llmErrors > 0) {\n    extras.push(`\u26A0\uFE0F ${llmErrors} LLM extraction failures`);\n  }\n  return extras.length > 0 ? `\\n${extras.join(' | ')}` : '';\n}\n\nfunction formatRedditOutput(\n  urls: string[],\n  processResult: PostProcessResult,\n  fetchComments: boolean,\n  totalBatches: number,\n  tokensPerUrl: number,\n  extraStatus: string,\n): string {\n  const batchHeader = formatBatchHeader({\n    title: `Reddit Posts`,\n    totalItems: urls.length,\n    successful: processResult.successful,\n    failed: processResult.failed,\n    ...(fetchComments ? { extras: { 'Words used': processResult.totalWordsUsed.toLocaleString() } } : {}),\n    tokensPerItem: tokensPerUrl,\n    batches: totalBatches,\n  });\n\n  let data = processResult.contents.join('\\n\\n---\\n\\n');\n\n  if (processResult.skippedUrls.length > 0) {\n    data += '\\n\\n---\\n\\n';\n    data += `**Word limit reached (${REDDIT.MAX_WORDS_TOTAL.toLocaleString()} words).** The following posts were not included:\\n`;\n    for (const url of processResult.skippedUrls) {\n      data += `- ${url}\\n`;\n    }\n    data += `\\nCall get-reddit-post again with just these URLs.`;\n  }\n\n  return formatSuccess({\n    title: `Reddit Posts Fetched (${processResult.successful}/${urls.length})`,\n    summary: batchHeader + extraStatus,\n    data,\n    nextSteps: [\n      processResult.successful > 0 ? 'web-search to verify claims from Reddit discussions' : null,\n      processResult.successful > 0 ? 'scrape-links on URLs referenced in comments' : null,\n      processResult.failed > 0 ? 'Retry failed URLs individually' : null,\n    ].filter(Boolean) as string[],\n  });\n}\n\nfunction formatGetRedditPostsError(error: unknown): string {\n  const structuredError = classifyError(error);\n  return formatError({\n    code: structuredError.code,\n    message: structuredError.message,\n    retryable: structuredError.retryable,\n    toolName: 'get-reddit-post',\n    howToFix: ['Verify REDDIT_CLIENT_ID and REDDIT_CLIENT_SECRET are set'],\n    alternatives: [\n      'web-search(keywords=[\"topic reddit discussion\", \"topic reddit recommendations\"]) \u2014 search for Reddit content via web search instead',\n      'scrape-links(urls=[...the Reddit URLs...], use_llm=true, what_to_extract=\"Extract post content | top comments | recommendations\") \u2014 scrape Reddit pages directly as a fallback',\n    ],\n  });\n}\n\nexport async function handleGetRedditPosts(\n  urls: string[],\n  clientId: string,\n  clientSecret: string,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<GetRedditPostOutput>> {\n  const startTime = Date.now();\n  try {\n    const validationError = validatePostCount(urls.length);\n    if (validationError) return toolFailure(validationError);\n\n    const totalBatches = Math.ceil(urls.length / REDDIT.BATCH_SIZE);\n\n    await reporter.log('info', `Fetching ${urls.length} Reddit post(s)`);\n    await reporter.progress(20, 100, 'Fetching Reddit posts');\n    const client = new RedditClient(clientId, clientSecret);\n    const batchResult = await client.batchGetPosts(urls, true);\n    await reporter.progress(55, 100, 'Formatting posts and comments');\n\n    const processResult = await fetchAndProcessPosts(batchResult.results);\n    await reporter.progress(85, 100, 'Building output');\n\n    const extraStatus = buildRedditStatusExtras(\n      batchResult.rateLimitHits, false, 0,\n    );\n    const content = formatRedditOutput(\n      urls, processResult, true, totalBatches, 0, extraStatus,\n    );\n\n    const executionTime = Date.now() - startTime;\n    return toolSuccess(content, {\n      content,\n      metadata: {\n        total_items: urls.length,\n        successful: processResult.successful,\n        failed: processResult.failed,\n        execution_time_ms: executionTime,\n        rate_limit_hits: batchResult.rateLimitHits,\n      },\n    });\n  } catch (error) {\n    return toolFailure(formatGetRedditPostsError(error));\n  }\n}\n\nexport function registerSearchRedditTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'search-reddit',\n      title: 'Search Reddit',\n      description:\n        'Search Google for Reddit posts matching up to 100 queries. Returns a flat list of unique Reddit URLs ready to pipe into get-reddit-post.',\n      schema: searchRedditParamsSchema,\n      outputSchema: searchRedditOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async ({ queries }, ctx) => {\n      if (!getCapabilities().search) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('search')));\n      }\n\n      const env = parseEnv();\n      const reporter = createToolReporter(ctx, 'search-reddit');\n      const result = await handleSearchReddit(queries, env.SEARCH_API_KEY!, reporter);\n\n      await reporter.progress(100, 100, result.isError ? 'Reddit search failed' : 'Reddit search complete');\n      return toToolResponse(result);\n    },\n  );\n}\n\nexport function registerGetRedditPostTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'get-reddit-post',\n      title: 'Get Reddit Post',\n      description:\n        'Fetch up to 100 Reddit posts with full threaded comment trees. Returns the raw post content and all comments with author, score, and OP markers.',\n      schema: getRedditPostParamsSchema,\n      outputSchema: getRedditPostOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async ({ urls }, ctx) => {\n      if (!getCapabilities().reddit) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('reddit')));\n      }\n\n      const env = parseEnv();\n      const reporter = createToolReporter(ctx, 'get-reddit-post');\n      const result = await handleGetRedditPosts(\n        urls,\n        env.REDDIT_CLIENT_ID!,\n        env.REDDIT_CLIENT_SECRET!,\n        reporter,\n      );\n\n      await reporter.progress(100, 100, result.isError ? 'Reddit fetch failed' : 'Reddit fetch complete');\n      return toToolResponse(result);\n    },\n  );\n}\n"],
+  "mappings": "AAOA;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AACP,SAAS,oBAAoB;AAC7B,SAAS,oBAAmD;AAC5D,SAAS,QAAQ,iBAAiB,sBAAsB,gBAAgB;AACxE,SAAS,qBAAqB;AAC9B;AAAA,EAEE;AAAA,EACA;AAAA,EACA;AAAA,OACK;AACP;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AAMP,SAAS,WAAW,MAAsB;AACxC,QAAM,QAAQ,KAAK,QAAQ,qBAAqB,EAAE;AAClD,SAAO,MAAM,MAAM,KAAK,EAAE,OAAO,OAAK,EAAE,SAAS,CAAC,EAAE;AACtD;AASA,SAAS,eAAe,UAAqB,UAA2C;AACtF,MAAI,KAAK;AACT,MAAI,YAAY;AAChB,MAAI,QAAQ;AAEZ,aAAW,KAAK,UAAU;AACxB,UAAM,SAAS,KAAK,OAAO,EAAE,KAAK;AAClC,UAAM,KAAK,EAAE,OAAO,cAAc;AAClC,UAAM,QAAQ,EAAE,SAAS,IAAI,IAAI,EAAE,KAAK,KAAK,GAAG,EAAE,KAAK;AACvD,UAAM,aAAa,GAAG,MAAM,SAAS,EAAE,MAAM,KAAK,EAAE,MAAM,KAAK;AAAA;AAC/D,UAAM,YAAY,EAAE,KAAK,MAAM,IAAI,EAAE,IAAI,UAAQ,GAAG,MAAM,KAAK,IAAI,EAAE,EAAE,KAAK,IAAI;AAChF,UAAM,YAAY,GAAG,UAAU,GAAG,SAAS;AAAA;AAAA;AAC3C,UAAM,eAAe,WAAW,SAAS;AAEzC,QAAI,YAAY,eAAe,YAAY,QAAQ,EAAG;AAEtD,UAAM;AACN,iBAAa;AACb;AAAA,EACF;AAEA,SAAO,EAAE,IAAI,WAAW,OAAO,WAAW,SAAS,SAAS,MAAM;AACpE;AASA,SAAS,WAAW,QAAoB,eAAwB,UAAuC;AACrG,QAAM,EAAE,MAAM,SAAS,IAAI;AAC3B,MAAI,KAAK,MAAM,KAAK,KAAK;AAAA;AAAA;AACzB,QAAM,OAAO,KAAK,SAAS,eAAU,KAAK,MAAM,wBAAS,KAAK,KAAK,qBAAS,KAAK,YAAY;AAAA;AAC7F,QAAM,aAAM,KAAK,GAAG;AAAA;AAAA;AAEpB,MAAI,YAAY,WAAW,EAAE;AAE7B,MAAI,KAAK,MAAM;AACb,UAAM,cAAc;AAAA;AAAA,EAAuB,KAAK,IAAI;AAAA;AAAA;AACpD,iBAAa,WAAW,WAAW;AACnC,UAAM;AAAA,EACR;AAEA,MAAI,gBAAgB;AACpB,MAAI,oBAAoB;AAExB,MAAI,iBAAiB,SAAS,SAAS,GAAG;AACxC,UAAM,iBAAiB,KAAK,IAAI,GAAG,WAAW,SAAS;AACvD,UAAM,iBAAiB,eAAe,UAAU,cAAc;AAC9D,oBAAgB,eAAe;AAC/B,wBAAoB,eAAe;AAEnC,UAAM,qBAAqB,eAAe,KAAK,IAAI,KAAK,YAAY,WAAW,eAAe,UAAU,eAAe,CAAC;AAAA;AAAA;AACxH,UAAM,eAAe;AACrB,iBAAa,eAAe;AAE5B,QAAI,eAAe,YAAY,GAAG;AAChC,YAAM;AAAA,GAAM,eAAe,SAAS;AAAA;AAAA;AAAA,IACtC;AAAA,EACF;AAEA,SAAO,EAAE,IAAI,WAAW,eAAe,kBAAkB;AAC3D;AAMA,eAAsB,mBACpB,SACA,QACA,WAAyB,eACyB;AAClD,MAAI;AACF,UAAM,YAAY,KAAK,IAAI;AAC3B,UAAM,UAAU,QAAQ,MAAM,GAAG,GAAG;AACpC,UAAM,SAAS,IAAI,aAAa,MAAM;AACtC,UAAM,SAAS,IAAI,QAAQ,yBAAyB,QAAQ,MAAM,UAAU;AAC5E,UAAM,SAAS,SAAS,IAAI,KAAK,kBAAkB;AACnD,UAAM,UAAU,MAAM,OAAO,qBAAqB,OAAO;AAGzD,UAAM,UAAU,oBAAI,IAAY;AAChC,eAAW,aAAa,QAAQ,OAAO,GAAG;AACxC,iBAAW,UAAU,WAAW;AAC9B,YAAI,OAAO,IAAK,SAAQ,IAAI,OAAO,GAAG;AAAA,MACxC;AAAA,IACF;AAEA,QAAI,QAAQ,SAAS,GAAG;AACtB,aAAO,YAAY,YAAY;AAAA,QAC7B,MAAM;AAAA,QACN,SAAS,uCAAuC,QAAQ,MAAM;AAAA,QAC9D,UAAU;AAAA,QACV,UAAU,CAAC,uCAAuC,gBAAgB;AAAA,QAClE,cAAc,CAAC,6FAAwF;AAAA,MACzG,CAAC,CAAC;AAAA,IACJ;AAEA,UAAM,UAAU,CAAC,GAAG,OAAO;AAC3B,UAAM,UAAU,QAAQ,KAAK,IAAI;AAEjC,UAAM,SAAS,IAAI,QAAQ,SAAS,QAAQ,MAAM,8BAA8B,QAAQ,MAAM,UAAU;AACxG,UAAM,SAAS,SAAS,KAAK,KAAK,wBAAwB;AAE1D,UAAM,gBAAgB,KAAK,IAAI,IAAI;AACnC,WAAO,YAAY,SAAS;AAAA,MAC1B;AAAA,MACA,UAAU;AAAA,QACR,aAAa,QAAQ;AAAA,QACrB,YAAY,QAAQ;AAAA,QACpB,QAAQ;AAAA,QACR,mBAAmB;AAAA,MACrB;AAAA,IACF,CAAC;AAAA,EACH,SAAS,OAAO;AACd,UAAM,kBAAkB,cAAc,KAAK;AAC3C,WAAO,YAAY,YAAY;AAAA,MAC7B,MAAM,gBAAgB;AAAA,MACtB,SAAS,gBAAgB;AAAA,MACzB,WAAW,gBAAgB;AAAA,MAC3B,UAAU;AAAA,MACV,UAAU,CAAC,wCAAwC;AAAA,IACrD,CAAC,CAAC;AAAA,EACJ;AACF;AAsBA,SAAS,kBAAkB,UAAiC;AAC1D,MAAI,WAAW,OAAO,WAAW;AAC/B,WAAO,YAAY;AAAA,MACjB,MAAM;AAAA,MACN,SAAS,WAAW,OAAO,SAAS,qCAAqC,QAAQ;AAAA,MACjF,UAAU;AAAA,MACV,UAAU,CAAC,gBAAgB,OAAO,YAAY,QAAQ,qBAAqB;AAAA,MAC3E,cAAc;AAAA,QACZ,iKAA4J,OAAO,SAAS;AAAA,MAC9K;AAAA,IACF,CAAC;AAAA,EACH;AACA,MAAI,WAAW,OAAO,WAAW;AAC/B,WAAO,YAAY;AAAA,MACjB,MAAM;AAAA,MACN,SAAS,WAAW,OAAO,SAAS,oCAAoC,QAAQ;AAAA,MAChF,UAAU;AAAA,MACV,UAAU,CAAC,UAAU,WAAW,OAAO,SAAS,mBAAmB;AAAA,IACrE,CAAC;AAAA,EACH;AACA,SAAO;AACT;AAEA,eAAe,qBACb,SAC4B;AAC5B,MAAI,SAAS;AACb,QAAM,iBAA2B,CAAC;AAClC,QAAM,kBAA4B,CAAC;AACnC,MAAI,aAAa;AACjB,MAAI,iBAAiB;AACrB,QAAM,cAAwB,CAAC;AAE/B,aAAW,CAAC,KAAK,MAAM,KAAK,SAAS;AACnC,QAAI,kBAAkB,OAAO;AAC3B;AACA,qBAAe,KAAK,qBAAgB,GAAG;AAAA;AAAA,GAAQ,OAAO,OAAO,GAAG;AAChE;AAAA,IACF;AAEA,QAAI,kBAAkB,OAAO,iBAAiB;AAC5C,kBAAY,KAAK,GAAG;AACpB;AAAA,IACF;AAEA,UAAM,YAAY,WAAW,QAAQ,MAAM,OAAO,kBAAkB;AACpE,sBAAkB,UAAU;AAC5B,oBAAgB,KAAK,UAAU,EAAE;AACjC;AAAA,EACF;AAEA,QAAM,WAAW,CAAC,GAAG,gBAAgB,GAAG,eAAe;AACvD,SAAO,EAAE,YAAY,QAAQ,WAAW,GAAG,cAAc,OAAO,UAAU,gBAAgB,YAAY;AACxG;AAEA,SAAS,wBACP,eACA,cACA,WACQ;AACR,QAAM,SAAmB,CAAC;AAC1B,MAAI,gBAAgB,EAAG,QAAO,KAAK,gBAAM,aAAa,qBAAqB;AAC3E,MAAI,CAAC,cAAc;AACjB,WAAO,KAAK,2FAA4E;AAAA,EAC1F,WAAW,YAAY,GAAG;AACxB,WAAO,KAAK,gBAAM,SAAS,0BAA0B;AAAA,EACvD;AACA,SAAO,OAAO,SAAS,IAAI;AAAA,EAAK,OAAO,KAAK,KAAK,CAAC,KAAK;AACzD;AAEA,SAAS,mBACP,MACA,eACA,eACA,cACA,cACA,aACQ;AACR,QAAM,cAAc,kBAAkB;AAAA,IACpC,OAAO;AAAA,IACP,YAAY,KAAK;AAAA,IACjB,YAAY,cAAc;AAAA,IAC1B,QAAQ,cAAc;AAAA,IACtB,GAAI,gBAAgB,EAAE,QAAQ,EAAE,cAAc,cAAc,eAAe,eAAe,EAAE,EAAE,IAAI,CAAC;AAAA,IACnG,eAAe;AAAA,IACf,SAAS;AAAA,EACX,CAAC;AAED,MAAI,OAAO,cAAc,SAAS,KAAK,aAAa;AAEpD,MAAI,cAAc,YAAY,SAAS,GAAG;AACxC,YAAQ;AACR,YAAQ,yBAAyB,OAAO,gBAAgB,eAAe,CAAC;AAAA;AACxE,eAAW,OAAO,cAAc,aAAa;AAC3C,cAAQ,KAAK,GAAG;AAAA;AAAA,IAClB;AACA,YAAQ;AAAA;AAAA,EACV;AAEA,SAAO,cAAc;AAAA,IACnB,OAAO,yBAAyB,cAAc,UAAU,IAAI,KAAK,MAAM;AAAA,IACvE,SAAS,cAAc;AAAA,IACvB;AAAA,IACA,WAAW;AAAA,MACT,cAAc,aAAa,IAAI,wDAAwD;AAAA,MACvF,cAAc,aAAa,IAAI,gDAAgD;AAAA,MAC/E,cAAc,SAAS,IAAI,mCAAmC;AAAA,IAChE,EAAE,OAAO,OAAO;AAAA,EAClB,CAAC;AACH;AAEA,SAAS,0BAA0B,OAAwB;AACzD,QAAM,kBAAkB,cAAc,KAAK;AAC3C,SAAO,YAAY;AAAA,IACjB,MAAM,gBAAgB;AAAA,IACtB,SAAS,gBAAgB;AAAA,IACzB,WAAW,gBAAgB;AAAA,IAC3B,UAAU;AAAA,IACV,UAAU,CAAC,0DAA0D;AAAA,IACrE,cAAc;AAAA,MACZ;AAAA,MACA;AAAA,IACF;AAAA,EACF,CAAC;AACH;AAEA,eAAsB,qBACpB,MACA,UACA,cACA,WAAyB,eAC0B;AACnD,QAAM,YAAY,KAAK,IAAI;AAC3B,MAAI;AACF,UAAM,kBAAkB,kBAAkB,KAAK,MAAM;AACrD,QAAI,gBAAiB,QAAO,YAAY,eAAe;AAEvD,UAAM,eAAe,KAAK,KAAK,KAAK,SAAS,OAAO,UAAU;AAE9D,UAAM,SAAS,IAAI,QAAQ,YAAY,KAAK,MAAM,iBAAiB;AACnE,UAAM,SAAS,SAAS,IAAI,KAAK,uBAAuB;AACxD,UAAM,SAAS,IAAI,aAAa,UAAU,YAAY;AACtD,UAAM,cAAc,MAAM,OAAO,cAAc,MAAM,IAAI;AACzD,UAAM,SAAS,SAAS,IAAI,KAAK,+BAA+B;AAEhE,UAAM,gBAAgB,MAAM,qBAAqB,YAAY,OAAO;AACpE,UAAM,SAAS,SAAS,IAAI,KAAK,iBAAiB;AAElD,UAAM,cAAc;AAAA,MAClB,YAAY;AAAA,MAAe;AAAA,MAAO;AAAA,IACpC;AACA,UAAM,UAAU;AAAA,MACd;AAAA,MAAM;AAAA,MAAe;AAAA,MAAM;AAAA,MAAc;AAAA,MAAG;AAAA,IAC9C;AAEA,UAAM,gBAAgB,KAAK,IAAI,IAAI;AACnC,WAAO,YAAY,SAAS;AAAA,MAC1B;AAAA,MACA,UAAU;AAAA,QACR,aAAa,KAAK;AAAA,QAClB,YAAY,cAAc;AAAA,QAC1B,QAAQ,cAAc;AAAA,QACtB,mBAAmB;AAAA,QACnB,iBAAiB,YAAY;AAAA,MAC/B;AAAA,IACF,CAAC;AAAA,EACH,SAAS,OAAO;AACd,WAAO,YAAY,0BAA0B,KAAK,CAAC;AAAA,EACrD;AACF;AAEO,SAAS,yBAAyB,QAAyB;AAChE,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,EAAE,QAAQ,GAAG,QAAQ;AAC1B,UAAI,CAAC,gBAAgB,EAAE,QAAQ;AAC7B,eAAO,eAAe,YAAY,qBAAqB,QAAQ,CAAC,CAAC;AAAA,MACnE;AAEA,YAAM,MAAM,SAAS;AACrB,YAAM,WAAW,mBAAmB,KAAK,eAAe;AACxD,YAAM,SAAS,MAAM,mBAAmB,SAAS,IAAI,gBAAiB,QAAQ;AAE9E,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,yBAAyB,wBAAwB;AACpG,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;AAEO,SAAS,0BAA0B,QAAyB;AACjE,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,EAAE,KAAK,GAAG,QAAQ;AACvB,UAAI,CAAC,gBAAgB,EAAE,QAAQ;AAC7B,eAAO,eAAe,YAAY,qBAAqB,QAAQ,CAAC,CAAC;AAAA,MACnE;AAEA,YAAM,MAAM,SAAS;AACrB,YAAM,WAAW,mBAAmB,KAAK,iBAAiB;AAC1D,YAAM,SAAS,MAAM;AAAA,QACnB;AAAA,QACA,IAAI;AAAA,QACJ,IAAI;AAAA,QACJ;AAAA,MACF;AAEA,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,wBAAwB,uBAAuB;AAClG,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;",
   "names": []
 }

package/dist/src/tools/scrape.js

@@ -157,14 +157,11 @@ ${content}`);
 }
 function buildScrapeMetadata(params, metrics, tokensPerUrl, totalBatches, executionTime) {
   return {
-    total_urls: params.urls.length,
+    total_items: params.urls.length,
     successful: metrics.successful,
     failed: metrics.failed,
-    total_credits: metrics.totalCredits,
     execution_time_ms: executionTime,
-    tokens_per_url: tokensPerUrl,
-    total_token_budget: TOKEN_BUDGETS.SCRAPER,
-    batches_processed: totalBatches
+    total_credits: metrics.totalCredits
   };
 }
 function buildScrapeResponse(params, contents, metrics, tokensPerUrl, totalBatches, llmErrors, executionTime) {
@@ -223,9 +220,9 @@ async function handleScrapeLinks(params, reporter = NOOP_REPORTER) {
       'search-reddit(queries=["topic discussion", "topic recommendations"]) \u2014 get community insights as an alternative'
     ]);
   }
-  const enhancedInstruction = enhanceExtractionInstruction(params.what_to_extract);
+  const enhancedInstruction = enhanceExtractionInstruction(params.extract);
   await reporter.progress(35, 100, "Fetching page content");
-  const results = await clients.client.scrapeMultiple(validUrls, { timeout: params.timeout });
+  const results = await clients.client.scrapeMultiple(validUrls, { timeout: 300 });
   mcpLog("info", `Scraping complete. Processing ${results.length} results...`, "scrape");
   await reporter.log("info", `Fetched ${results.length} scrape response(s) from the provider`);
   await reporter.progress(60, 100, "Cleaning and classifying scrape results");
@@ -262,7 +259,7 @@ function registerScrapeLinksTool(server) {
     {
       name: "scrape-links",
       title: "Scrape Links",
-      description: 'Scrape 1-50 web pages and run LLM extraction on each. Provide what_to_extract with specific targets (e.g., "Extract pricing tiers | feature limits | API rate limits"). Token budget (32K) is split across URLs: 3 URLs get ~10K tokens each (deep), 10 get ~3K (balanced), 50 get ~640 (scan).',
+      description: "Scrape up to 100 web pages and run LLM extraction on each. Returns only the data you specify in the extract field \u2014 everything else is filtered out.",
       schema: scrapeLinksParamsSchema,
       outputSchema: scrapeLinksOutputSchema,
       annotations: {

package/dist/src/tools/scrape.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../src/tools/scrape.ts"],
-  "sourcesContent": ["/**\n * Scrape Links Tool Handler\n * Implements robust error handling that NEVER crashes the MCP server\n */\n\nimport type { MCPServer } from 'mcp-use/server';\n\nimport { SCRAPER, CONCURRENCY, getCapabilities, getMissingEnvMessage } from '../config/index.js';\nimport {\n  scrapeLinksOutputSchema,\n  scrapeLinksParamsSchema,\n  type ScrapeLinksParams,\n  type ScrapeLinksOutput,\n} from '../schemas/scrape-links.js';\nimport { ScraperClient } from '../clients/scraper.js';\nimport { MarkdownCleaner } from '../services/markdown-cleaner.js';\nimport { createLLMProcessor, processContentWithLLM } from '../services/llm-processor.js';\nimport { removeMetaTags } from '../utils/markdown-formatter.js';\nimport { classifyError } from '../utils/errors.js';\nimport { pMap } from '../utils/concurrency.js';\nimport {\n  mcpLog,\n  formatSuccess,\n  formatError,\n  formatBatchHeader,\n  formatDuration,\n  TOKEN_BUDGETS,\n  calculateTokenAllocation,\n} from './utils.js';\nimport {\n  createToolReporter,\n  NOOP_REPORTER,\n  toolFailure,\n  toolSuccess,\n  toToolResponse,\n  type ToolExecutionResult,\n  type ToolReporter,\n} from './mcp-helpers.js';\n\n// Module-level singleton - MarkdownCleaner is stateless\nconst markdownCleaner = new MarkdownCleaner();\n\n// Extraction prefix/suffix are kept in runtime config to avoid YAML indirection.\nfunction getExtractionPrefix(): string {\n  return SCRAPER.EXTRACTION_PREFIX;\n}\n\nfunction getExtractionSuffix(): string {\n  return SCRAPER.EXTRACTION_SUFFIX;\n}\n\nfunction enhanceExtractionInstruction(instruction: string | undefined): string {\n  const base = instruction || 'Extract the main content and key information from this page.';\n  return `${getExtractionPrefix()}\\n\\n${base}\\n\\n${getExtractionSuffix()}`;\n}\n\n// --- Internal types for decomposed helpers ---\n\ninterface ProcessedResult {\n  url: string;\n  content: string;\n  index: number;\n}\n\ninterface ScrapeMetrics {\n  successful: number;\n  failed: number;\n  totalCredits: number;\n}\n\ninterface ScrapePhaseResult {\n  successItems: ProcessedResult[];\n  failedContents: string[];\n  metrics: ScrapeMetrics;\n}\n\ninterface ScrapeClients {\n  client: ScraperClient;\n  llmProcessor: ReturnType<typeof createLLMProcessor>;\n}\n\n// --- Helpers ---\n\nfunction createScrapeErrorResponse(\n  code: string,\n  message: string,\n  startTime: number,\n  totalUrls: number,\n  retryable = false,\n  alternatives?: string[],\n): ToolExecutionResult<ScrapeLinksOutput> {\n  return toolFailure(\n    `${formatError({\n      code,\n      message,\n      retryable,\n      toolName: 'scrape-links',\n      howToFix: code === 'NO_URLS' ? ['Provide at least one valid URL'] : undefined,\n      alternatives,\n    })}\\n\\nExecution time: ${formatDuration(Date.now() - startTime)}`,\n  );\n}\n\nfunction validateAndPartitionUrls(urls: string[]): { validUrls: string[]; invalidUrls: string[] } {\n  const validUrls: string[] = [];\n  const invalidUrls: string[] = [];\n  for (const url of urls) {\n    try {\n      new URL(url);\n      validUrls.push(url);\n    } catch {\n      invalidUrls.push(url);\n    }\n  }\n  return { validUrls, invalidUrls };\n}\n\nfunction initializeScrapeClients(): ScrapeClients {\n  const client = new ScraperClient();\n  const llmProcessor = createLLMProcessor();\n  return { client, llmProcessor };\n}\n\nfunction processScrapeResults(\n  results: Awaited<ReturnType<ScraperClient['scrapeMultiple']>>,\n  invalidUrls: string[],\n): ScrapePhaseResult {\n  const successItems: ProcessedResult[] = [];\n  const failedContents: string[] = [];\n  let successful = 0;\n  let failed = 0;\n  let totalCredits = 0;\n\n  for (const invalidUrl of invalidUrls) {\n    failed++;\n    failedContents.push(`## ${invalidUrl}\\n\\n\u274C Invalid URL format`);\n  }\n\n  for (let i = 0; i < results.length; i++) {\n    const result = results[i];\n    if (!result) {\n      failed++;\n      failedContents.push(`## Unknown URL\\n\\n\u274C No result returned`);\n      continue;\n    }\n\n    mcpLog('debug', `[${i + 1}/${results.length}] Processing ${result.url}`, 'scrape');\n\n    if (result.error || result.statusCode < 200 || result.statusCode >= 300) {\n      failed++;\n      const errorMsg = result.error?.message || result.content || `HTTP ${result.statusCode}`;\n      failedContents.push(`## ${result.url}\\n\\n\u274C Failed to scrape: ${errorMsg}`);\n      mcpLog('warning', `[${i + 1}/${results.length}] Failed: ${errorMsg}`, 'scrape');\n      continue;\n    }\n\n    successful++;\n    totalCredits += result.credits;\n\n    let content: string;\n    try {\n      content = markdownCleaner.processContent(result.content);\n    } catch {\n      content = result.content;\n    }\n\n    successItems.push({ url: result.url, content, index: i });\n  }\n\n  return { successItems, failedContents, metrics: { successful, failed, totalCredits } };\n}\n\nasync function processItemsWithLlm(\n  successItems: ProcessedResult[],\n  enhancedInstruction: string,\n  tokensPerUrl: number,\n  llmProcessor: ReturnType<typeof createLLMProcessor>,\n): Promise<{ items: ProcessedResult[]; llmErrors: number }> {\n  let llmErrors = 0;\n\n  if (!llmProcessor || successItems.length === 0) {\n    if (!llmProcessor && successItems.length > 0) {\n      mcpLog('warning', 'LLM unavailable (LLM_EXTRACTION_API_KEY not set). Returning raw scraped content.', 'scrape');\n    }\n    return { items: successItems, llmErrors };\n  }\n\n  mcpLog('info', `Starting parallel LLM extraction for ${successItems.length} pages (concurrency: ${CONCURRENCY.LLM_EXTRACTION})`, 'scrape');\n\n  const llmResults = await pMap(successItems, async (item) => {\n    mcpLog('debug', `LLM extracting ${item.url} (${tokensPerUrl} tokens)...`, 'scrape');\n\n    const llmResult = await processContentWithLLM(\n      item.content,\n      { use_llm: true, what_to_extract: enhancedInstruction, max_tokens: tokensPerUrl },\n      llmProcessor,\n    );\n\n    if (llmResult.processed) {\n      mcpLog('debug', `LLM extraction complete for ${item.url}`, 'scrape');\n      return { ...item, content: llmResult.content };\n    }\n\n    llmErrors++;\n    mcpLog('warning', `LLM extraction failed for ${item.url}: ${llmResult.error || 'unknown reason'}`, 'scrape');\n    return item;\n  }, CONCURRENCY.LLM_EXTRACTION);\n\n  return { items: llmResults, llmErrors };\n}\n\nfunction assembleContentEntries(successItems: ProcessedResult[], failedContents: string[]): string[] {\n  const contents = [...failedContents];\n  for (const item of successItems) {\n    let content = item.content;\n    try {\n      content = removeMetaTags(content);\n    } catch {\n      // Use content as-is\n    }\n    contents.push(`## ${item.url}\\n\\n${content}`);\n  }\n  return contents;\n}\n\nfunction buildScrapeMetadata(\n  params: ScrapeLinksParams,\n  metrics: ScrapeMetrics,\n  tokensPerUrl: number,\n  totalBatches: number,\n  executionTime: number,\n): ScrapeLinksOutput['metadata'] {\n  return {\n    total_urls: params.urls.length,\n    successful: metrics.successful,\n    failed: metrics.failed,\n    total_credits: metrics.totalCredits,\n    execution_time_ms: executionTime,\n    tokens_per_url: tokensPerUrl,\n    total_token_budget: TOKEN_BUDGETS.SCRAPER,\n    batches_processed: totalBatches,\n  };\n}\n\nfunction buildScrapeResponse(\n  params: ScrapeLinksParams,\n  contents: string[],\n  metrics: ScrapeMetrics,\n  tokensPerUrl: number,\n  totalBatches: number,\n  llmErrors: number,\n  executionTime: number,\n): { content: string; structuredContent: ScrapeLinksOutput } {\n  const batchHeader = formatBatchHeader({\n    title: `Scraped Content (${params.urls.length} URLs)`,\n    totalItems: params.urls.length,\n    successful: metrics.successful,\n    failed: metrics.failed,\n    tokensPerItem: tokensPerUrl,\n    batches: totalBatches,\n    extras: {\n      'Credits used': metrics.totalCredits,\n      ...(llmErrors > 0 ? { 'LLM extraction failures': llmErrors } : {}),\n    },\n  });\n\n  const formattedContent = formatSuccess({\n    title: 'Scraping Complete',\n    summary: batchHeader,\n    data: contents.join('\\n\\n---\\n\\n'),\n    nextSteps: [\n      metrics.successful > 0 ? 'web-search or search-reddit to cross-check claims from scraped content' : null,\n      metrics.failed > 0 ? 'Retry failed URLs with timeout=60' : null,\n    ].filter(Boolean) as string[],\n    metadata: {\n      'Execution time': formatDuration(executionTime),\n      'Token budget': TOKEN_BUDGETS.SCRAPER.toLocaleString(),\n    },\n  });\n\n  const metadata = buildScrapeMetadata(params, metrics, tokensPerUrl, totalBatches, executionTime);\n  return { content: formattedContent, structuredContent: { content: formattedContent, metadata } };\n}\n\n/**\n * Handle scrape links request\n * NEVER throws - always returns a valid response with content and metadata\n */\nexport async function handleScrapeLinks(\n  params: ScrapeLinksParams,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<ScrapeLinksOutput>> {\n  const startTime = Date.now();\n\n  if (!params.urls || params.urls.length === 0) {\n    return createScrapeErrorResponse('NO_URLS', 'No URLs provided', startTime, params.urls?.length || 0);\n  }\n\n  const { validUrls, invalidUrls } = validateAndPartitionUrls(params.urls);\n  await reporter.log('info', `Validated ${validUrls.length} scrapeable URL(s) and ${invalidUrls.length} invalid URL(s)`);\n\n  if (validUrls.length === 0) {\n    return createScrapeErrorResponse('INVALID_URLS', `All ${params.urls.length} URLs are invalid`, startTime, params.urls.length, false, [\n      'web-search(keywords=[\"topic documentation\", \"topic guide\"]) \u2014 search for valid URLs first, then scrape the results',\n      'search-reddit(queries=[\"topic recommendations\"]) \u2014 find discussion URLs to scrape instead',\n    ]);\n  }\n\n  const tokensPerUrl = calculateTokenAllocation(validUrls.length, TOKEN_BUDGETS.SCRAPER);\n  const totalBatches = Math.ceil(validUrls.length / SCRAPER.BATCH_SIZE);\n\n  mcpLog('info', `Starting scrape: ${validUrls.length} URL(s), ${tokensPerUrl} tokens/URL, ${totalBatches} batch(es)`, 'scrape');\n  await reporter.progress(15, 100, 'Preparing scraper clients');\n\n  let clients: ScrapeClients;\n  try {\n    clients = initializeScrapeClients();\n  } catch (error) {\n    const err = classifyError(error);\n    return createScrapeErrorResponse('CLIENT_INIT_FAILED', `Failed to initialize scraper: ${err.message}`, startTime, params.urls.length, false, [\n      'web-search(keywords=[\"topic key findings\", \"topic summary\", \"topic overview\"]) \u2014 search for information instead of scraping',\n      'search-reddit(queries=[\"topic discussion\", \"topic recommendations\"]) \u2014 get community insights as an alternative',\n    ]);\n  }\n\n  const enhancedInstruction = enhanceExtractionInstruction(params.what_to_extract);\n\n  await reporter.progress(35, 100, 'Fetching page content');\n  const results = await clients.client.scrapeMultiple(validUrls, { timeout: params.timeout });\n  mcpLog('info', `Scraping complete. Processing ${results.length} results...`, 'scrape');\n  await reporter.log('info', `Fetched ${results.length} scrape response(s) from the provider`);\n  await reporter.progress(60, 100, 'Cleaning and classifying scrape results');\n\n  const { successItems, failedContents, metrics } = processScrapeResults(results, invalidUrls);\n\n  if (successItems.length > 0) {\n    await reporter.progress(80, 100, 'Running LLM extraction over scraped pages');\n  }\n  const { items: processedItems, llmErrors } = await processItemsWithLlm(\n    successItems, enhancedInstruction, tokensPerUrl, clients.llmProcessor,\n  );\n\n  const contents = assembleContentEntries(processedItems, failedContents);\n  const executionTime = Date.now() - startTime;\n\n  mcpLog('info', `Completed: ${metrics.successful} successful, ${metrics.failed} failed, ${metrics.totalCredits} credits used`, 'scrape');\n  await reporter.log(\n    'info',\n    `Scrape completed with ${metrics.successful} success(es), ${metrics.failed} failure(s), and ${llmErrors} LLM extraction issue(s)`,\n  );\n\n  const result = buildScrapeResponse(\n    params,\n    contents,\n    metrics,\n    tokensPerUrl,\n    totalBatches,\n    llmErrors,\n    executionTime,\n  );\n  return toolSuccess(result.content, result.structuredContent);\n}\n\nexport function registerScrapeLinksTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'scrape-links',\n      title: 'Scrape Links',\n      description:\n        'Scrape 1-50 web pages and run LLM extraction on each. Provide what_to_extract with specific targets (e.g., \"Extract pricing tiers | feature limits | API rate limits\"). Token budget (32K) is split across URLs: 3 URLs get ~10K tokens each (deep), 10 get ~3K (balanced), 50 get ~640 (scan).',\n      schema: scrapeLinksParamsSchema,\n      outputSchema: scrapeLinksOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async (args, ctx) => {\n      if (!getCapabilities().scraping) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('scraping')));\n      }\n\n      const reporter = createToolReporter(ctx, 'scrape-links');\n      const result = await handleScrapeLinks(args, reporter);\n\n      await reporter.progress(100, 100, result.isError ? 'Scrape failed' : 'Scrape complete');\n      return toToolResponse(result);\n    },\n  );\n}\n"],
-  "mappings": "AAOA,SAAS,SAAS,aAAa,iBAAiB,4BAA4B;AAC5E;AAAA,EACE;AAAA,EACA;AAAA,OAGK;AACP,SAAS,qBAAqB;AAC9B,SAAS,uBAAuB;AAChC,SAAS,oBAAoB,6BAA6B;AAC1D,SAAS,sBAAsB;AAC/B,SAAS,qBAAqB;AAC9B,SAAS,YAAY;AACrB;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OACK;AACP;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AAGP,MAAM,kBAAkB,IAAI,gBAAgB;AAG5C,SAAS,sBAA8B;AACrC,SAAO,QAAQ;AACjB;AAEA,SAAS,sBAA8B;AACrC,SAAO,QAAQ;AACjB;AAEA,SAAS,6BAA6B,aAAyC;AAC7E,QAAM,OAAO,eAAe;AAC5B,SAAO,GAAG,oBAAoB,CAAC;AAAA;AAAA,EAAO,IAAI;AAAA;AAAA,EAAO,oBAAoB,CAAC;AACxE;AA6BA,SAAS,0BACP,MACA,SACA,WACA,WACA,YAAY,OACZ,cACwC;AACxC,SAAO;AAAA,IACL,GAAG,YAAY;AAAA,MACb;AAAA,MACA;AAAA,MACA;AAAA,MACA,UAAU;AAAA,MACV,UAAU,SAAS,YAAY,CAAC,gCAAgC,IAAI;AAAA,MACpE;AAAA,IACF,CAAC,CAAC;AAAA;AAAA,kBAAuB,eAAe,KAAK,IAAI,IAAI,SAAS,CAAC;AAAA,EACjE;AACF;AAEA,SAAS,yBAAyB,MAAgE;AAChG,QAAM,YAAsB,CAAC;AAC7B,QAAM,cAAwB,CAAC;AAC/B,aAAW,OAAO,MAAM;AACtB,QAAI;AACF,UAAI,IAAI,GAAG;AACX,gBAAU,KAAK,GAAG;AAAA,IACpB,QAAQ;AACN,kBAAY,KAAK,GAAG;AAAA,IACtB;AAAA,EACF;AACA,SAAO,EAAE,WAAW,YAAY;AAClC;AAEA,SAAS,0BAAyC;AAChD,QAAM,SAAS,IAAI,cAAc;AACjC,QAAM,eAAe,mBAAmB;AACxC,SAAO,EAAE,QAAQ,aAAa;AAChC;AAEA,SAAS,qBACP,SACA,aACmB;AACnB,QAAM,eAAkC,CAAC;AACzC,QAAM,iBAA2B,CAAC;AAClC,MAAI,aAAa;AACjB,MAAI,SAAS;AACb,MAAI,eAAe;AAEnB,aAAW,cAAc,aAAa;AACpC;AACA,mBAAe,KAAK,MAAM,UAAU;AAAA;AAAA,0BAA0B;AAAA,EAChE;AAEA,WAAS,IAAI,GAAG,IAAI,QAAQ,QAAQ,KAAK;AACvC,UAAM,SAAS,QAAQ,CAAC;AACxB,QAAI,CAAC,QAAQ;AACX;AACA,qBAAe,KAAK;AAAA;AAAA,0BAAwC;AAC5D;AAAA,IACF;AAEA,WAAO,SAAS,IAAI,IAAI,CAAC,IAAI,QAAQ,MAAM,gBAAgB,OAAO,GAAG,IAAI,QAAQ;AAEjF,QAAI,OAAO,SAAS,OAAO,aAAa,OAAO,OAAO,cAAc,KAAK;AACvE;AACA,YAAM,WAAW,OAAO,OAAO,WAAW,OAAO,WAAW,QAAQ,OAAO,UAAU;AACrF,qBAAe,KAAK,MAAM,OAAO,GAAG;AAAA;AAAA,2BAA2B,QAAQ,EAAE;AACzE,aAAO,WAAW,IAAI,IAAI,CAAC,IAAI,QAAQ,MAAM,aAAa,QAAQ,IAAI,QAAQ;AAC9E;AAAA,IACF;AAEA;AACA,oBAAgB,OAAO;AAEvB,QAAI;AACJ,QAAI;AACF,gBAAU,gBAAgB,eAAe,OAAO,OAAO;AAAA,IACzD,QAAQ;AACN,gBAAU,OAAO;AAAA,IACnB;AAEA,iBAAa,KAAK,EAAE,KAAK,OAAO,KAAK,SAAS,OAAO,EAAE,CAAC;AAAA,EAC1D;AAEA,SAAO,EAAE,cAAc,gBAAgB,SAAS,EAAE,YAAY,QAAQ,aAAa,EAAE;AACvF;AAEA,eAAe,oBACb,cACA,qBACA,cACA,cAC0D;AAC1D,MAAI,YAAY;AAEhB,MAAI,CAAC,gBAAgB,aAAa,WAAW,GAAG;AAC9C,QAAI,CAAC,gBAAgB,aAAa,SAAS,GAAG;AAC5C,aAAO,WAAW,oFAAoF,QAAQ;AAAA,IAChH;AACA,WAAO,EAAE,OAAO,cAAc,UAAU;AAAA,EAC1C;AAEA,SAAO,QAAQ,wCAAwC,aAAa,MAAM,wBAAwB,YAAY,cAAc,KAAK,QAAQ;AAEzI,QAAM,aAAa,MAAM,KAAK,cAAc,OAAO,SAAS;AAC1D,WAAO,SAAS,kBAAkB,KAAK,GAAG,KAAK,YAAY,eAAe,QAAQ;AAElF,UAAM,YAAY,MAAM;AAAA,MACtB,KAAK;AAAA,MACL,EAAE,SAAS,MAAM,iBAAiB,qBAAqB,YAAY,aAAa;AAAA,MAChF;AAAA,IACF;AAEA,QAAI,UAAU,WAAW;AACvB,aAAO,SAAS,+BAA+B,KAAK,GAAG,IAAI,QAAQ;AACnE,aAAO,EAAE,GAAG,MAAM,SAAS,UAAU,QAAQ;AAAA,IAC/C;AAEA;AACA,WAAO,WAAW,6BAA6B,KAAK,GAAG,KAAK,UAAU,SAAS,gBAAgB,IAAI,QAAQ;AAC3G,WAAO;AAAA,EACT,GAAG,YAAY,cAAc;AAE7B,SAAO,EAAE,OAAO,YAAY,UAAU;AACxC;AAEA,SAAS,uBAAuB,cAAiC,gBAAoC;AACnG,QAAM,WAAW,CAAC,GAAG,cAAc;AACnC,aAAW,QAAQ,cAAc;AAC/B,QAAI,UAAU,KAAK;AACnB,QAAI;AACF,gBAAU,eAAe,OAAO;AAAA,IAClC,QAAQ;AAAA,IAER;AACA,aAAS,KAAK,MAAM,KAAK,GAAG;AAAA;AAAA,EAAO,OAAO,EAAE;AAAA,EAC9C;AACA,SAAO;AACT;AAEA,SAAS,oBACP,QACA,SACA,cACA,cACA,eAC+B;AAC/B,SAAO;AAAA,IACL,YAAY,OAAO,KAAK;AAAA,IACxB,YAAY,QAAQ;AAAA,IACpB,QAAQ,QAAQ;AAAA,IAChB,eAAe,QAAQ;AAAA,IACvB,mBAAmB;AAAA,IACnB,gBAAgB;AAAA,IAChB,oBAAoB,cAAc;AAAA,IAClC,mBAAmB;AAAA,EACrB;AACF;AAEA,SAAS,oBACP,QACA,UACA,SACA,cACA,cACA,WACA,eAC2D;AAC3D,QAAM,cAAc,kBAAkB;AAAA,IACpC,OAAO,oBAAoB,OAAO,KAAK,MAAM;AAAA,IAC7C,YAAY,OAAO,KAAK;AAAA,IACxB,YAAY,QAAQ;AAAA,IACpB,QAAQ,QAAQ;AAAA,IAChB,eAAe;AAAA,IACf,SAAS;AAAA,IACT,QAAQ;AAAA,MACN,gBAAgB,QAAQ;AAAA,MACxB,GAAI,YAAY,IAAI,EAAE,2BAA2B,UAAU,IAAI,CAAC;AAAA,IAClE;AAAA,EACF,CAAC;AAED,QAAM,mBAAmB,cAAc;AAAA,IACrC,OAAO;AAAA,IACP,SAAS;AAAA,IACT,MAAM,SAAS,KAAK,aAAa;AAAA,IACjC,WAAW;AAAA,MACT,QAAQ,aAAa,IAAI,2EAA2E;AAAA,MACpG,QAAQ,SAAS,IAAI,sCAAsC;AAAA,IAC7D,EAAE,OAAO,OAAO;AAAA,IAChB,UAAU;AAAA,MACR,kBAAkB,eAAe,aAAa;AAAA,MAC9C,gBAAgB,cAAc,QAAQ,eAAe;AAAA,IACvD;AAAA,EACF,CAAC;AAED,QAAM,WAAW,oBAAoB,QAAQ,SAAS,cAAc,cAAc,aAAa;AAC/F,SAAO,EAAE,SAAS,kBAAkB,mBAAmB,EAAE,SAAS,kBAAkB,SAAS,EAAE;AACjG;AAMA,eAAsB,kBACpB,QACA,WAAyB,eACwB;AACjD,QAAM,YAAY,KAAK,IAAI;AAE3B,MAAI,CAAC,OAAO,QAAQ,OAAO,KAAK,WAAW,GAAG;AAC5C,WAAO,0BAA0B,WAAW,oBAAoB,WAAW,OAAO,MAAM,UAAU,CAAC;AAAA,EACrG;AAEA,QAAM,EAAE,WAAW,YAAY,IAAI,yBAAyB,OAAO,IAAI;AACvE,QAAM,SAAS,IAAI,QAAQ,aAAa,UAAU,MAAM,0BAA0B,YAAY,MAAM,iBAAiB;AAErH,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,0BAA0B,gBAAgB,OAAO,OAAO,KAAK,MAAM,qBAAqB,WAAW,OAAO,KAAK,QAAQ,OAAO;AAAA,MACnI;AAAA,MACA;AAAA,IACF,CAAC;AAAA,EACH;AAEA,QAAM,eAAe,yBAAyB,UAAU,QAAQ,cAAc,OAAO;AACrF,QAAM,eAAe,KAAK,KAAK,UAAU,SAAS,QAAQ,UAAU;AAEpE,SAAO,QAAQ,oBAAoB,UAAU,MAAM,YAAY,YAAY,gBAAgB,YAAY,cAAc,QAAQ;AAC7H,QAAM,SAAS,SAAS,IAAI,KAAK,2BAA2B;AAE5D,MAAI;AACJ,MAAI;AACF,cAAU,wBAAwB;AAAA,EACpC,SAAS,OAAO;AACd,UAAM,MAAM,cAAc,KAAK;AAC/B,WAAO,0BAA0B,sBAAsB,iCAAiC,IAAI,OAAO,IAAI,WAAW,OAAO,KAAK,QAAQ,OAAO;AAAA,MAC3I;AAAA,MACA;AAAA,IACF,CAAC;AAAA,EACH;AAEA,QAAM,sBAAsB,6BAA6B,OAAO,eAAe;AAE/E,QAAM,SAAS,SAAS,IAAI,KAAK,uBAAuB;AACxD,QAAM,UAAU,MAAM,QAAQ,OAAO,eAAe,WAAW,EAAE,SAAS,OAAO,QAAQ,CAAC;AAC1F,SAAO,QAAQ,iCAAiC,QAAQ,MAAM,eAAe,QAAQ;AACrF,QAAM,SAAS,IAAI,QAAQ,WAAW,QAAQ,MAAM,uCAAuC;AAC3F,QAAM,SAAS,SAAS,IAAI,KAAK,yCAAyC;AAE1E,QAAM,EAAE,cAAc,gBAAgB,QAAQ,IAAI,qBAAqB,SAAS,WAAW;AAE3F,MAAI,aAAa,SAAS,GAAG;AAC3B,UAAM,SAAS,SAAS,IAAI,KAAK,2CAA2C;AAAA,EAC9E;AACA,QAAM,EAAE,OAAO,gBAAgB,UAAU,IAAI,MAAM;AAAA,IACjD;AAAA,IAAc;AAAA,IAAqB;AAAA,IAAc,QAAQ;AAAA,EAC3D;AAEA,QAAM,WAAW,uBAAuB,gBAAgB,cAAc;AACtE,QAAM,gBAAgB,KAAK,IAAI,IAAI;AAEnC,SAAO,QAAQ,cAAc,QAAQ,UAAU,gBAAgB,QAAQ,MAAM,YAAY,QAAQ,YAAY,iBAAiB,QAAQ;AACtI,QAAM,SAAS;AAAA,IACb;AAAA,IACA,yBAAyB,QAAQ,UAAU,iBAAiB,QAAQ,MAAM,oBAAoB,SAAS;AAAA,EACzG;AAEA,QAAM,SAAS;AAAA,IACb;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,EACF;AACA,SAAO,YAAY,OAAO,SAAS,OAAO,iBAAiB;AAC7D;AAEO,SAAS,wBAAwB,QAAyB;AAC/D,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,MAAM,QAAQ;AACnB,UAAI,CAAC,gBAAgB,EAAE,UAAU;AAC/B,eAAO,eAAe,YAAY,qBAAqB,UAAU,CAAC,CAAC;AAAA,MACrE;AAEA,YAAM,WAAW,mBAAmB,KAAK,cAAc;AACvD,YAAM,SAAS,MAAM,kBAAkB,MAAM,QAAQ;AAErD,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,kBAAkB,iBAAiB;AACtF,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;",
+  "sourcesContent": ["/**\n * Scrape Links Tool Handler\n * Implements robust error handling that NEVER crashes the MCP server\n */\n\nimport type { MCPServer } from 'mcp-use/server';\n\nimport { SCRAPER, CONCURRENCY, getCapabilities, getMissingEnvMessage } from '../config/index.js';\nimport {\n  scrapeLinksOutputSchema,\n  scrapeLinksParamsSchema,\n  type ScrapeLinksParams,\n  type ScrapeLinksOutput,\n} from '../schemas/scrape-links.js';\nimport { ScraperClient } from '../clients/scraper.js';\nimport { MarkdownCleaner } from '../services/markdown-cleaner.js';\nimport { createLLMProcessor, processContentWithLLM } from '../services/llm-processor.js';\nimport { removeMetaTags } from '../utils/markdown-formatter.js';\nimport { classifyError } from '../utils/errors.js';\nimport { pMap } from '../utils/concurrency.js';\nimport {\n  mcpLog,\n  formatSuccess,\n  formatError,\n  formatBatchHeader,\n  formatDuration,\n  TOKEN_BUDGETS,\n  calculateTokenAllocation,\n} from './utils.js';\nimport {\n  createToolReporter,\n  NOOP_REPORTER,\n  toolFailure,\n  toolSuccess,\n  toToolResponse,\n  type ToolExecutionResult,\n  type ToolReporter,\n} from './mcp-helpers.js';\n\n// Module-level singleton - MarkdownCleaner is stateless\nconst markdownCleaner = new MarkdownCleaner();\n\n// Extraction prefix/suffix are kept in runtime config to avoid YAML indirection.\nfunction getExtractionPrefix(): string {\n  return SCRAPER.EXTRACTION_PREFIX;\n}\n\nfunction getExtractionSuffix(): string {\n  return SCRAPER.EXTRACTION_SUFFIX;\n}\n\nfunction enhanceExtractionInstruction(instruction: string | undefined): string {\n  const base = instruction || 'Extract the main content and key information from this page.';\n  return `${getExtractionPrefix()}\\n\\n${base}\\n\\n${getExtractionSuffix()}`;\n}\n\n// --- Internal types for decomposed helpers ---\n\ninterface ProcessedResult {\n  url: string;\n  content: string;\n  index: number;\n}\n\ninterface ScrapeMetrics {\n  successful: number;\n  failed: number;\n  totalCredits: number;\n}\n\ninterface ScrapePhaseResult {\n  successItems: ProcessedResult[];\n  failedContents: string[];\n  metrics: ScrapeMetrics;\n}\n\ninterface ScrapeClients {\n  client: ScraperClient;\n  llmProcessor: ReturnType<typeof createLLMProcessor>;\n}\n\n// --- Helpers ---\n\nfunction createScrapeErrorResponse(\n  code: string,\n  message: string,\n  startTime: number,\n  totalUrls: number,\n  retryable = false,\n  alternatives?: string[],\n): ToolExecutionResult<ScrapeLinksOutput> {\n  return toolFailure(\n    `${formatError({\n      code,\n      message,\n      retryable,\n      toolName: 'scrape-links',\n      howToFix: code === 'NO_URLS' ? ['Provide at least one valid URL'] : undefined,\n      alternatives,\n    })}\\n\\nExecution time: ${formatDuration(Date.now() - startTime)}`,\n  );\n}\n\nfunction validateAndPartitionUrls(urls: string[]): { validUrls: string[]; invalidUrls: string[] } {\n  const validUrls: string[] = [];\n  const invalidUrls: string[] = [];\n  for (const url of urls) {\n    try {\n      new URL(url);\n      validUrls.push(url);\n    } catch {\n      invalidUrls.push(url);\n    }\n  }\n  return { validUrls, invalidUrls };\n}\n\nfunction initializeScrapeClients(): ScrapeClients {\n  const client = new ScraperClient();\n  const llmProcessor = createLLMProcessor();\n  return { client, llmProcessor };\n}\n\nfunction processScrapeResults(\n  results: Awaited<ReturnType<ScraperClient['scrapeMultiple']>>,\n  invalidUrls: string[],\n): ScrapePhaseResult {\n  const successItems: ProcessedResult[] = [];\n  const failedContents: string[] = [];\n  let successful = 0;\n  let failed = 0;\n  let totalCredits = 0;\n\n  for (const invalidUrl of invalidUrls) {\n    failed++;\n    failedContents.push(`## ${invalidUrl}\\n\\n\u274C Invalid URL format`);\n  }\n\n  for (let i = 0; i < results.length; i++) {\n    const result = results[i];\n    if (!result) {\n      failed++;\n      failedContents.push(`## Unknown URL\\n\\n\u274C No result returned`);\n      continue;\n    }\n\n    mcpLog('debug', `[${i + 1}/${results.length}] Processing ${result.url}`, 'scrape');\n\n    if (result.error || result.statusCode < 200 || result.statusCode >= 300) {\n      failed++;\n      const errorMsg = result.error?.message || result.content || `HTTP ${result.statusCode}`;\n      failedContents.push(`## ${result.url}\\n\\n\u274C Failed to scrape: ${errorMsg}`);\n      mcpLog('warning', `[${i + 1}/${results.length}] Failed: ${errorMsg}`, 'scrape');\n      continue;\n    }\n\n    successful++;\n    totalCredits += result.credits;\n\n    let content: string;\n    try {\n      content = markdownCleaner.processContent(result.content);\n    } catch {\n      content = result.content;\n    }\n\n    successItems.push({ url: result.url, content, index: i });\n  }\n\n  return { successItems, failedContents, metrics: { successful, failed, totalCredits } };\n}\n\nasync function processItemsWithLlm(\n  successItems: ProcessedResult[],\n  enhancedInstruction: string,\n  tokensPerUrl: number,\n  llmProcessor: ReturnType<typeof createLLMProcessor>,\n): Promise<{ items: ProcessedResult[]; llmErrors: number }> {\n  let llmErrors = 0;\n\n  if (!llmProcessor || successItems.length === 0) {\n    if (!llmProcessor && successItems.length > 0) {\n      mcpLog('warning', 'LLM unavailable (LLM_EXTRACTION_API_KEY not set). Returning raw scraped content.', 'scrape');\n    }\n    return { items: successItems, llmErrors };\n  }\n\n  mcpLog('info', `Starting parallel LLM extraction for ${successItems.length} pages (concurrency: ${CONCURRENCY.LLM_EXTRACTION})`, 'scrape');\n\n  const llmResults = await pMap(successItems, async (item) => {\n    mcpLog('debug', `LLM extracting ${item.url} (${tokensPerUrl} tokens)...`, 'scrape');\n\n    const llmResult = await processContentWithLLM(\n      item.content,\n      { use_llm: true, what_to_extract: enhancedInstruction, max_tokens: tokensPerUrl },\n      llmProcessor,\n    );\n\n    if (llmResult.processed) {\n      mcpLog('debug', `LLM extraction complete for ${item.url}`, 'scrape');\n      return { ...item, content: llmResult.content };\n    }\n\n    llmErrors++;\n    mcpLog('warning', `LLM extraction failed for ${item.url}: ${llmResult.error || 'unknown reason'}`, 'scrape');\n    return item;\n  }, CONCURRENCY.LLM_EXTRACTION);\n\n  return { items: llmResults, llmErrors };\n}\n\nfunction assembleContentEntries(successItems: ProcessedResult[], failedContents: string[]): string[] {\n  const contents = [...failedContents];\n  for (const item of successItems) {\n    let content = item.content;\n    try {\n      content = removeMetaTags(content);\n    } catch {\n      // Use content as-is\n    }\n    contents.push(`## ${item.url}\\n\\n${content}`);\n  }\n  return contents;\n}\n\nfunction buildScrapeMetadata(\n  params: ScrapeLinksParams,\n  metrics: ScrapeMetrics,\n  tokensPerUrl: number,\n  totalBatches: number,\n  executionTime: number,\n): ScrapeLinksOutput['metadata'] {\n  return {\n    total_items: params.urls.length,\n    successful: metrics.successful,\n    failed: metrics.failed,\n    execution_time_ms: executionTime,\n    total_credits: metrics.totalCredits,\n  };\n}\n\nfunction buildScrapeResponse(\n  params: ScrapeLinksParams,\n  contents: string[],\n  metrics: ScrapeMetrics,\n  tokensPerUrl: number,\n  totalBatches: number,\n  llmErrors: number,\n  executionTime: number,\n): { content: string; structuredContent: ScrapeLinksOutput } {\n  const batchHeader = formatBatchHeader({\n    title: `Scraped Content (${params.urls.length} URLs)`,\n    totalItems: params.urls.length,\n    successful: metrics.successful,\n    failed: metrics.failed,\n    tokensPerItem: tokensPerUrl,\n    batches: totalBatches,\n    extras: {\n      'Credits used': metrics.totalCredits,\n      ...(llmErrors > 0 ? { 'LLM extraction failures': llmErrors } : {}),\n    },\n  });\n\n  const formattedContent = formatSuccess({\n    title: 'Scraping Complete',\n    summary: batchHeader,\n    data: contents.join('\\n\\n---\\n\\n'),\n    nextSteps: [\n      metrics.successful > 0 ? 'web-search or search-reddit to cross-check claims from scraped content' : null,\n      metrics.failed > 0 ? 'Retry failed URLs with timeout=60' : null,\n    ].filter(Boolean) as string[],\n    metadata: {\n      'Execution time': formatDuration(executionTime),\n      'Token budget': TOKEN_BUDGETS.SCRAPER.toLocaleString(),\n    },\n  });\n\n  const metadata = buildScrapeMetadata(params, metrics, tokensPerUrl, totalBatches, executionTime);\n  return { content: formattedContent, structuredContent: { content: formattedContent, metadata } };\n}\n\n/**\n * Handle scrape links request\n * NEVER throws - always returns a valid response with content and metadata\n */\nexport async function handleScrapeLinks(\n  params: ScrapeLinksParams,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<ScrapeLinksOutput>> {\n  const startTime = Date.now();\n\n  if (!params.urls || params.urls.length === 0) {\n    return createScrapeErrorResponse('NO_URLS', 'No URLs provided', startTime, params.urls?.length || 0);\n  }\n\n  const { validUrls, invalidUrls } = validateAndPartitionUrls(params.urls);\n  await reporter.log('info', `Validated ${validUrls.length} scrapeable URL(s) and ${invalidUrls.length} invalid URL(s)`);\n\n  if (validUrls.length === 0) {\n    return createScrapeErrorResponse('INVALID_URLS', `All ${params.urls.length} URLs are invalid`, startTime, params.urls.length, false, [\n      'web-search(keywords=[\"topic documentation\", \"topic guide\"]) \u2014 search for valid URLs first, then scrape the results',\n      'search-reddit(queries=[\"topic recommendations\"]) \u2014 find discussion URLs to scrape instead',\n    ]);\n  }\n\n  const tokensPerUrl = calculateTokenAllocation(validUrls.length, TOKEN_BUDGETS.SCRAPER);\n  const totalBatches = Math.ceil(validUrls.length / SCRAPER.BATCH_SIZE);\n\n  mcpLog('info', `Starting scrape: ${validUrls.length} URL(s), ${tokensPerUrl} tokens/URL, ${totalBatches} batch(es)`, 'scrape');\n  await reporter.progress(15, 100, 'Preparing scraper clients');\n\n  let clients: ScrapeClients;\n  try {\n    clients = initializeScrapeClients();\n  } catch (error) {\n    const err = classifyError(error);\n    return createScrapeErrorResponse('CLIENT_INIT_FAILED', `Failed to initialize scraper: ${err.message}`, startTime, params.urls.length, false, [\n      'web-search(keywords=[\"topic key findings\", \"topic summary\", \"topic overview\"]) \u2014 search for information instead of scraping',\n      'search-reddit(queries=[\"topic discussion\", \"topic recommendations\"]) \u2014 get community insights as an alternative',\n    ]);\n  }\n\n  const enhancedInstruction = enhanceExtractionInstruction(params.extract);\n\n  await reporter.progress(35, 100, 'Fetching page content');\n  const results = await clients.client.scrapeMultiple(validUrls, { timeout: 300 });\n  mcpLog('info', `Scraping complete. Processing ${results.length} results...`, 'scrape');\n  await reporter.log('info', `Fetched ${results.length} scrape response(s) from the provider`);\n  await reporter.progress(60, 100, 'Cleaning and classifying scrape results');\n\n  const { successItems, failedContents, metrics } = processScrapeResults(results, invalidUrls);\n\n  if (successItems.length > 0) {\n    await reporter.progress(80, 100, 'Running LLM extraction over scraped pages');\n  }\n  const { items: processedItems, llmErrors } = await processItemsWithLlm(\n    successItems, enhancedInstruction, tokensPerUrl, clients.llmProcessor,\n  );\n\n  const contents = assembleContentEntries(processedItems, failedContents);\n  const executionTime = Date.now() - startTime;\n\n  mcpLog('info', `Completed: ${metrics.successful} successful, ${metrics.failed} failed, ${metrics.totalCredits} credits used`, 'scrape');\n  await reporter.log(\n    'info',\n    `Scrape completed with ${metrics.successful} success(es), ${metrics.failed} failure(s), and ${llmErrors} LLM extraction issue(s)`,\n  );\n\n  const result = buildScrapeResponse(\n    params,\n    contents,\n    metrics,\n    tokensPerUrl,\n    totalBatches,\n    llmErrors,\n    executionTime,\n  );\n  return toolSuccess(result.content, result.structuredContent);\n}\n\nexport function registerScrapeLinksTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'scrape-links',\n      title: 'Scrape Links',\n      description:\n        'Scrape up to 100 web pages and run LLM extraction on each. Returns only the data you specify in the extract field \u2014 everything else is filtered out.',\n      schema: scrapeLinksParamsSchema,\n      outputSchema: scrapeLinksOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async (args, ctx) => {\n      if (!getCapabilities().scraping) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('scraping')));\n      }\n\n      const reporter = createToolReporter(ctx, 'scrape-links');\n      const result = await handleScrapeLinks(args, reporter);\n\n      await reporter.progress(100, 100, result.isError ? 'Scrape failed' : 'Scrape complete');\n      return toToolResponse(result);\n    },\n  );\n}\n"],
+  "mappings": "AAOA,SAAS,SAAS,aAAa,iBAAiB,4BAA4B;AAC5E;AAAA,EACE;AAAA,EACA;AAAA,OAGK;AACP,SAAS,qBAAqB;AAC9B,SAAS,uBAAuB;AAChC,SAAS,oBAAoB,6BAA6B;AAC1D,SAAS,sBAAsB;AAC/B,SAAS,qBAAqB;AAC9B,SAAS,YAAY;AACrB;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OACK;AACP;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AAGP,MAAM,kBAAkB,IAAI,gBAAgB;AAG5C,SAAS,sBAA8B;AACrC,SAAO,QAAQ;AACjB;AAEA,SAAS,sBAA8B;AACrC,SAAO,QAAQ;AACjB;AAEA,SAAS,6BAA6B,aAAyC;AAC7E,QAAM,OAAO,eAAe;AAC5B,SAAO,GAAG,oBAAoB,CAAC;AAAA;AAAA,EAAO,IAAI;AAAA;AAAA,EAAO,oBAAoB,CAAC;AACxE;AA6BA,SAAS,0BACP,MACA,SACA,WACA,WACA,YAAY,OACZ,cACwC;AACxC,SAAO;AAAA,IACL,GAAG,YAAY;AAAA,MACb;AAAA,MACA;AAAA,MACA;AAAA,MACA,UAAU;AAAA,MACV,UAAU,SAAS,YAAY,CAAC,gCAAgC,IAAI;AAAA,MACpE;AAAA,IACF,CAAC,CAAC;AAAA;AAAA,kBAAuB,eAAe,KAAK,IAAI,IAAI,SAAS,CAAC;AAAA,EACjE;AACF;AAEA,SAAS,yBAAyB,MAAgE;AAChG,QAAM,YAAsB,CAAC;AAC7B,QAAM,cAAwB,CAAC;AAC/B,aAAW,OAAO,MAAM;AACtB,QAAI;AACF,UAAI,IAAI,GAAG;AACX,gBAAU,KAAK,GAAG;AAAA,IACpB,QAAQ;AACN,kBAAY,KAAK,GAAG;AAAA,IACtB;AAAA,EACF;AACA,SAAO,EAAE,WAAW,YAAY;AAClC;AAEA,SAAS,0BAAyC;AAChD,QAAM,SAAS,IAAI,cAAc;AACjC,QAAM,eAAe,mBAAmB;AACxC,SAAO,EAAE,QAAQ,aAAa;AAChC;AAEA,SAAS,qBACP,SACA,aACmB;AACnB,QAAM,eAAkC,CAAC;AACzC,QAAM,iBAA2B,CAAC;AAClC,MAAI,aAAa;AACjB,MAAI,SAAS;AACb,MAAI,eAAe;AAEnB,aAAW,cAAc,aAAa;AACpC;AACA,mBAAe,KAAK,MAAM,UAAU;AAAA;AAAA,0BAA0B;AAAA,EAChE;AAEA,WAAS,IAAI,GAAG,IAAI,QAAQ,QAAQ,KAAK;AACvC,UAAM,SAAS,QAAQ,CAAC;AACxB,QAAI,CAAC,QAAQ;AACX;AACA,qBAAe,KAAK;AAAA;AAAA,0BAAwC;AAC5D;AAAA,IACF;AAEA,WAAO,SAAS,IAAI,IAAI,CAAC,IAAI,QAAQ,MAAM,gBAAgB,OAAO,GAAG,IAAI,QAAQ;AAEjF,QAAI,OAAO,SAAS,OAAO,aAAa,OAAO,OAAO,cAAc,KAAK;AACvE;AACA,YAAM,WAAW,OAAO,OAAO,WAAW,OAAO,WAAW,QAAQ,OAAO,UAAU;AACrF,qBAAe,KAAK,MAAM,OAAO,GAAG;AAAA;AAAA,2BAA2B,QAAQ,EAAE;AACzE,aAAO,WAAW,IAAI,IAAI,CAAC,IAAI,QAAQ,MAAM,aAAa,QAAQ,IAAI,QAAQ;AAC9E;AAAA,IACF;AAEA;AACA,oBAAgB,OAAO;AAEvB,QAAI;AACJ,QAAI;AACF,gBAAU,gBAAgB,eAAe,OAAO,OAAO;AAAA,IACzD,QAAQ;AACN,gBAAU,OAAO;AAAA,IACnB;AAEA,iBAAa,KAAK,EAAE,KAAK,OAAO,KAAK,SAAS,OAAO,EAAE,CAAC;AAAA,EAC1D;AAEA,SAAO,EAAE,cAAc,gBAAgB,SAAS,EAAE,YAAY,QAAQ,aAAa,EAAE;AACvF;AAEA,eAAe,oBACb,cACA,qBACA,cACA,cAC0D;AAC1D,MAAI,YAAY;AAEhB,MAAI,CAAC,gBAAgB,aAAa,WAAW,GAAG;AAC9C,QAAI,CAAC,gBAAgB,aAAa,SAAS,GAAG;AAC5C,aAAO,WAAW,oFAAoF,QAAQ;AAAA,IAChH;AACA,WAAO,EAAE,OAAO,cAAc,UAAU;AAAA,EAC1C;AAEA,SAAO,QAAQ,wCAAwC,aAAa,MAAM,wBAAwB,YAAY,cAAc,KAAK,QAAQ;AAEzI,QAAM,aAAa,MAAM,KAAK,cAAc,OAAO,SAAS;AAC1D,WAAO,SAAS,kBAAkB,KAAK,GAAG,KAAK,YAAY,eAAe,QAAQ;AAElF,UAAM,YAAY,MAAM;AAAA,MACtB,KAAK;AAAA,MACL,EAAE,SAAS,MAAM,iBAAiB,qBAAqB,YAAY,aAAa;AAAA,MAChF;AAAA,IACF;AAEA,QAAI,UAAU,WAAW;AACvB,aAAO,SAAS,+BAA+B,KAAK,GAAG,IAAI,QAAQ;AACnE,aAAO,EAAE,GAAG,MAAM,SAAS,UAAU,QAAQ;AAAA,IAC/C;AAEA;AACA,WAAO,WAAW,6BAA6B,KAAK,GAAG,KAAK,UAAU,SAAS,gBAAgB,IAAI,QAAQ;AAC3G,WAAO;AAAA,EACT,GAAG,YAAY,cAAc;AAE7B,SAAO,EAAE,OAAO,YAAY,UAAU;AACxC;AAEA,SAAS,uBAAuB,cAAiC,gBAAoC;AACnG,QAAM,WAAW,CAAC,GAAG,cAAc;AACnC,aAAW,QAAQ,cAAc;AAC/B,QAAI,UAAU,KAAK;AACnB,QAAI;AACF,gBAAU,eAAe,OAAO;AAAA,IAClC,QAAQ;AAAA,IAER;AACA,aAAS,KAAK,MAAM,KAAK,GAAG;AAAA;AAAA,EAAO,OAAO,EAAE;AAAA,EAC9C;AACA,SAAO;AACT;AAEA,SAAS,oBACP,QACA,SACA,cACA,cACA,eAC+B;AAC/B,SAAO;AAAA,IACL,aAAa,OAAO,KAAK;AAAA,IACzB,YAAY,QAAQ;AAAA,IACpB,QAAQ,QAAQ;AAAA,IAChB,mBAAmB;AAAA,IACnB,eAAe,QAAQ;AAAA,EACzB;AACF;AAEA,SAAS,oBACP,QACA,UACA,SACA,cACA,cACA,WACA,eAC2D;AAC3D,QAAM,cAAc,kBAAkB;AAAA,IACpC,OAAO,oBAAoB,OAAO,KAAK,MAAM;AAAA,IAC7C,YAAY,OAAO,KAAK;AAAA,IACxB,YAAY,QAAQ;AAAA,IACpB,QAAQ,QAAQ;AAAA,IAChB,eAAe;AAAA,IACf,SAAS;AAAA,IACT,QAAQ;AAAA,MACN,gBAAgB,QAAQ;AAAA,MACxB,GAAI,YAAY,IAAI,EAAE,2BAA2B,UAAU,IAAI,CAAC;AAAA,IAClE;AAAA,EACF,CAAC;AAED,QAAM,mBAAmB,cAAc;AAAA,IACrC,OAAO;AAAA,IACP,SAAS;AAAA,IACT,MAAM,SAAS,KAAK,aAAa;AAAA,IACjC,WAAW;AAAA,MACT,QAAQ,aAAa,IAAI,2EAA2E;AAAA,MACpG,QAAQ,SAAS,IAAI,sCAAsC;AAAA,IAC7D,EAAE,OAAO,OAAO;AAAA,IAChB,UAAU;AAAA,MACR,kBAAkB,eAAe,aAAa;AAAA,MAC9C,gBAAgB,cAAc,QAAQ,eAAe;AAAA,IACvD;AAAA,EACF,CAAC;AAED,QAAM,WAAW,oBAAoB,QAAQ,SAAS,cAAc,cAAc,aAAa;AAC/F,SAAO,EAAE,SAAS,kBAAkB,mBAAmB,EAAE,SAAS,kBAAkB,SAAS,EAAE;AACjG;AAMA,eAAsB,kBACpB,QACA,WAAyB,eACwB;AACjD,QAAM,YAAY,KAAK,IAAI;AAE3B,MAAI,CAAC,OAAO,QAAQ,OAAO,KAAK,WAAW,GAAG;AAC5C,WAAO,0BAA0B,WAAW,oBAAoB,WAAW,OAAO,MAAM,UAAU,CAAC;AAAA,EACrG;AAEA,QAAM,EAAE,WAAW,YAAY,IAAI,yBAAyB,OAAO,IAAI;AACvE,QAAM,SAAS,IAAI,QAAQ,aAAa,UAAU,MAAM,0BAA0B,YAAY,MAAM,iBAAiB;AAErH,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,0BAA0B,gBAAgB,OAAO,OAAO,KAAK,MAAM,qBAAqB,WAAW,OAAO,KAAK,QAAQ,OAAO;AAAA,MACnI;AAAA,MACA;AAAA,IACF,CAAC;AAAA,EACH;AAEA,QAAM,eAAe,yBAAyB,UAAU,QAAQ,cAAc,OAAO;AACrF,QAAM,eAAe,KAAK,KAAK,UAAU,SAAS,QAAQ,UAAU;AAEpE,SAAO,QAAQ,oBAAoB,UAAU,MAAM,YAAY,YAAY,gBAAgB,YAAY,cAAc,QAAQ;AAC7H,QAAM,SAAS,SAAS,IAAI,KAAK,2BAA2B;AAE5D,MAAI;AACJ,MAAI;AACF,cAAU,wBAAwB;AAAA,EACpC,SAAS,OAAO;AACd,UAAM,MAAM,cAAc,KAAK;AAC/B,WAAO,0BAA0B,sBAAsB,iCAAiC,IAAI,OAAO,IAAI,WAAW,OAAO,KAAK,QAAQ,OAAO;AAAA,MAC3I;AAAA,MACA;AAAA,IACF,CAAC;AAAA,EACH;AAEA,QAAM,sBAAsB,6BAA6B,OAAO,OAAO;AAEvE,QAAM,SAAS,SAAS,IAAI,KAAK,uBAAuB;AACxD,QAAM,UAAU,MAAM,QAAQ,OAAO,eAAe,WAAW,EAAE,SAAS,IAAI,CAAC;AAC/E,SAAO,QAAQ,iCAAiC,QAAQ,MAAM,eAAe,QAAQ;AACrF,QAAM,SAAS,IAAI,QAAQ,WAAW,QAAQ,MAAM,uCAAuC;AAC3F,QAAM,SAAS,SAAS,IAAI,KAAK,yCAAyC;AAE1E,QAAM,EAAE,cAAc,gBAAgB,QAAQ,IAAI,qBAAqB,SAAS,WAAW;AAE3F,MAAI,aAAa,SAAS,GAAG;AAC3B,UAAM,SAAS,SAAS,IAAI,KAAK,2CAA2C;AAAA,EAC9E;AACA,QAAM,EAAE,OAAO,gBAAgB,UAAU,IAAI,MAAM;AAAA,IACjD;AAAA,IAAc;AAAA,IAAqB;AAAA,IAAc,QAAQ;AAAA,EAC3D;AAEA,QAAM,WAAW,uBAAuB,gBAAgB,cAAc;AACtE,QAAM,gBAAgB,KAAK,IAAI,IAAI;AAEnC,SAAO,QAAQ,cAAc,QAAQ,UAAU,gBAAgB,QAAQ,MAAM,YAAY,QAAQ,YAAY,iBAAiB,QAAQ;AACtI,QAAM,SAAS;AAAA,IACb;AAAA,IACA,yBAAyB,QAAQ,UAAU,iBAAiB,QAAQ,MAAM,oBAAoB,SAAS;AAAA,EACzG;AAEA,QAAM,SAAS;AAAA,IACb;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,EACF;AACA,SAAO,YAAY,OAAO,SAAS,OAAO,iBAAiB;AAC7D;AAEO,SAAS,wBAAwB,QAAyB;AAC/D,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,MAAM,QAAQ;AACnB,UAAI,CAAC,gBAAgB,EAAE,UAAU;AAC/B,eAAO,eAAe,YAAY,qBAAqB,UAAU,CAAC,CAAC;AAAA,MACrE;AAEA,YAAM,WAAW,mBAAmB,KAAK,cAAc;AACvD,YAAM,SAAS,MAAM,kBAAkB,MAAM,QAAQ;AAErD,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,kBAAkB,iBAAiB;AACtF,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;",
   "names": []
 }

package/dist/src/tools/search.js

@@ -41,7 +41,7 @@ function buildRawOutput(keywords, aggregation, searches) {
     aggregation.thresholdNote
   );
 }
-function buildClassifiedOutput(classification, aggregation, objective, totalKeywords) {
+function buildClassifiedOutput(classification, aggregation, extract, totalKeywords) {
   const rankedUrls = aggregation.rankedUrls;
   const urlByRank = new Map(rankedUrls.map((u) => [u.rank, u]));
   const tiers = {
@@ -62,7 +62,7 @@ function buildClassifiedOutput(classification, aggregation, objective, totalKeyw
   }
   const lines = [];
   lines.push(`## ${classification.title}`);
-  lines.push(`> Objective: ${objective}`);
+  lines.push(`> Looking for: ${extract}`);
   lines.push(`> ${totalKeywords} queries \u2192 ${rankedUrls.length} URLs \u2192 ${tiers.high.length} highly relevant, ${tiers.maybe.length} possibly relevant`);
   lines.push("");
   lines.push(`**Summary:** ${classification.synthesis}`);
@@ -120,12 +120,10 @@ function buildMetadata(aggregation, executionTime, totalKeywords, searches, llmC
   });
   const lowYieldKeywords = searches.filter((s) => s.results.length <= 1).map((s) => s.keyword);
   return {
-    total_keywords: totalKeywords,
-    total_results: aggregation.rankedUrls.length,
+    total_items: totalKeywords,
+    successful: aggregation.rankedUrls.length,
+    failed: totalKeywords - searches.filter((s) => s.results.length > 0).length,
     execution_time_ms: executionTime,
-    total_unique_urls: aggregation.totalUniqueUrls,
-    consensus_url_count: aggregation.rankedUrls.filter((u) => u.isConsensus).length,
-    frequency_threshold: aggregation.frequencyThreshold,
     llm_classified: llmClassified,
     ...llmError ? { llm_error: llmError } : {},
     coverage_summary: coverageSummary,
@@ -151,16 +149,16 @@ function buildWebSearchError(error, params, startTime) {
     `${errorContent}
 
 Execution time: ${formatDuration(executionTime)}
-Keywords: ${params.keywords.length}`
+Queries: ${params.queries.length}`
   );
 }
 async function handleWebSearch(params, reporter = NOOP_REPORTER) {
   const startTime = Date.now();
   try {
-    mcpLog("info", `Searching for ${params.keywords.length} keyword(s)`, "search");
-    await reporter.log("info", `Searching for ${params.keywords.length} keyword(s)`);
+    mcpLog("info", `Searching for ${params.queries.length} query/queries`, "search");
+    await reporter.log("info", `Searching for ${params.queries.length} query/queries`);
     await reporter.progress(15, 100, "Submitting search queries");
-    const response = await executeSearches(params.keywords);
+    const response = await executeSearches(params.queries);
     await reporter.progress(50, 100, "Collected search results");
     const { aggregation } = processResults(response);
     await reporter.log(
@@ -177,13 +175,13 @@ async function handleWebSearch(params, reporter = NOOP_REPORTER) {
         llmError = "LLM unavailable (LLM_EXTRACTION_API_KEY not set). Falling back to raw output.";
         mcpLog("warning", llmError, "search");
       }
-      markdown = buildRawOutput(params.keywords, aggregation, response.searches);
+      markdown = buildRawOutput(params.queries, aggregation, response.searches);
       await reporter.progress(80, 100, "Ranking search results");
     } else {
       await reporter.progress(65, 100, "Classifying results by relevance");
       const classification = await classifySearchResults(
         aggregation.rankedUrls,
-        params.objective,
+        params.extract,
         response.totalKeywords,
         llmProcessor
       );
@@ -191,7 +189,7 @@ async function handleWebSearch(params, reporter = NOOP_REPORTER) {
         markdown = buildClassifiedOutput(
           classification.result,
           aggregation,
-          params.objective,
+          params.extract,
           response.totalKeywords
         );
         llmClassified = true;
@@ -199,7 +197,7 @@ async function handleWebSearch(params, reporter = NOOP_REPORTER) {
       } else {
         llmError = classification.error ?? "Unknown classification error";
         mcpLog("warning", `Classification failed, falling back to raw: ${llmError}`, "search");
-        markdown = buildRawOutput(params.keywords, aggregation, response.searches);
+        markdown = buildRawOutput(params.queries, aggregation, response.searches);
         await reporter.progress(85, 100, "Classification failed, using raw output");
       }
     }
@@ -216,7 +214,7 @@ async function handleWebSearch(params, reporter = NOOP_REPORTER) {
     await reporter.log("info", `Search completed with ${aggregation.rankedUrls.length} URLs (classified: ${llmClassified})`);
     const footer = `
 ---
-*${formatDuration(executionTime)} | ${aggregation.totalUniqueUrls} unique URLs | ${metadata.consensus_url_count} consensus | threshold \u2265${aggregation.frequencyThreshold}${llmClassified ? " | LLM classified" : ""}*`;
+*${formatDuration(executionTime)} | ${aggregation.totalUniqueUrls} unique URLs${llmClassified ? " | LLM classified" : ""}*`;
     const fullMarkdown = markdown + footer;
     return toolSuccess(fullMarkdown, { content: fullMarkdown, metadata });
   } catch (error) {
@@ -228,7 +226,7 @@ function registerWebSearchTool(server) {
     {
       name: "web-search",
       title: "Web Search",
-      description: "Search Google with 1-100 keywords in parallel. Each keyword runs as a separate search; results are aggregated, deduplicated, and ranked. Provide an objective describing what you need \u2014 an LLM classifies every result into 3 tiers (highly relevant / maybe relevant / other) based on titles and snippets alone (no URL fetching). Output: synthesis paragraph + tiered markdown table. Set raw=true to skip classification and get the traditional consensus-ranked URL list.",
+      description: "Run up to 100 Google searches in parallel, aggregate and deduplicate results, then classify each URL by relevance to your extract goal. Returns a tiered table: highly relevant, maybe relevant, and other. Set raw=true for unclassified ranked results.",
       schema: webSearchParamsSchema,
       outputSchema: webSearchOutputSchema,
       annotations: {

package/dist/src/tools/search.js.map

@@ -1,7 +1,7 @@
 {
   "version": 3,
   "sources": ["../../../src/tools/search.ts"],
-  "sourcesContent": ["/**\n * Web Search Tool Handler\n * NEVER throws - always returns structured response for graceful degradation\n */\n\nimport type { MCPServer } from 'mcp-use/server';\n\nimport { getCapabilities, getMissingEnvMessage } from '../config/index.js';\nimport {\n  webSearchOutputSchema,\n  webSearchParamsSchema,\n  type WebSearchParams,\n  type WebSearchOutput,\n} from '../schemas/web-search.js';\nimport { SearchClient } from '../clients/search.js';\nimport {\n  aggregateAndRank,\n  generateUnifiedOutput,\n} from '../utils/url-aggregator.js';\nimport { createLLMProcessor, classifySearchResults, type ClassificationResult } from '../services/llm-processor.js';\nimport { classifyError } from '../utils/errors.js';\nimport {\n  mcpLog,\n  formatError,\n  formatDuration,\n} from './utils.js';\nimport {\n  createToolReporter,\n  NOOP_REPORTER,\n  toolFailure,\n  toolSuccess,\n  toToolResponse,\n  type ToolExecutionResult,\n  type ToolReporter,\n} from './mcp-helpers.js';\n\n// --- Internal types ---\n\ninterface SearchAggregation {\n  readonly rankedUrls: ReturnType<typeof aggregateAndRank>['rankedUrls'];\n  readonly totalUniqueUrls: number;\n  readonly frequencyThreshold: number;\n  readonly thresholdNote?: string;\n}\n\ninterface SearchResponse {\n  searches: Parameters<typeof aggregateAndRank>[0];\n  totalKeywords: number;\n}\n\n// --- Helpers ---\n\nasync function executeSearches(keywords: string[]): Promise<SearchResponse> {\n  const client = new SearchClient();\n  return client.searchMultiple(keywords);\n}\n\nfunction processResults(response: SearchResponse): {\n  aggregation: SearchAggregation;\n  consensusUrls: SearchAggregation['rankedUrls'];\n} {\n  const aggregation = aggregateAndRank(response.searches, 5);\n  const consensusUrls = aggregation.rankedUrls.filter(u => u.isConsensus);\n  return { aggregation, consensusUrls };\n}\n\n// --- Raw output (traditional unified ranked list) ---\n\nfunction buildRawOutput(\n  keywords: string[],\n  aggregation: SearchAggregation,\n  searches: SearchResponse['searches'],\n): string {\n  return generateUnifiedOutput(\n    aggregation.rankedUrls, keywords, searches,\n    aggregation.totalUniqueUrls,\n    aggregation.frequencyThreshold, aggregation.thresholdNote,\n  );\n}\n\n// --- Classified output (3-tier LLM-classified table) ---\n\nfunction buildClassifiedOutput(\n  classification: ClassificationResult,\n  aggregation: SearchAggregation,\n  objective: string,\n  totalKeywords: number,\n): string {\n  const rankedUrls = aggregation.rankedUrls;\n\n  // Build lookup from rank \u2192 url data\n  const urlByRank = new Map(rankedUrls.map(u => [u.rank, u]));\n\n  // Build tier \u2192 entries mapping\n  const tiers = {\n    high: [] as typeof rankedUrls,\n    maybe: [] as typeof rankedUrls,\n    other: [] as typeof rankedUrls,\n  };\n\n  // Classify based on LLM response\n  const tierMap = new Map(classification.results.map(r => [r.rank, r.tier]));\n\n  for (const url of rankedUrls) {\n    const tier = tierMap.get(url.rank);\n    if (tier === 'HIGHLY_RELEVANT') {\n      tiers.high.push(url);\n    } else if (tier === 'MAYBE_RELEVANT') {\n      tiers.maybe.push(url);\n    } else {\n      tiers.other.push(url);\n    }\n  }\n\n  const lines: string[] = [];\n\n  // Header with generated title and synthesis\n  lines.push(`## ${classification.title}`);\n  lines.push(`> Objective: ${objective}`);\n  lines.push(`> ${totalKeywords} queries \u2192 ${rankedUrls.length} URLs \u2192 ${tiers.high.length} highly relevant, ${tiers.maybe.length} possibly relevant`);\n  lines.push('');\n  lines.push(`**Summary:** ${classification.synthesis}`);\n  lines.push('');\n\n  // Highly Relevant tier\n  if (tiers.high.length > 0) {\n    lines.push(`### Highly Relevant (${tiers.high.length})`);\n    lines.push('| # | URL | Seen in |');\n    lines.push('|---|-----|---------|');\n    for (const url of tiers.high) {\n      const coveragePct = Math.round(url.coverageRatio * 100);\n      const queries = url.queries.map(q => `\"${q}\"`).join(', ');\n      lines.push(`| ${url.rank} | [${url.title}](${url.url}) | ${url.frequency}/${totalKeywords} (${coveragePct}%) |`);\n    }\n    lines.push('');\n  }\n\n  // Maybe Relevant tier\n  if (tiers.maybe.length > 0) {\n    lines.push(`### Maybe Relevant (${tiers.maybe.length})`);\n    lines.push('| # | URL | Seen in |');\n    lines.push('|---|-----|---------|');\n    for (const url of tiers.maybe) {\n      const coveragePct = Math.round(url.coverageRatio * 100);\n      lines.push(`| ${url.rank} | [${url.title}](${url.url}) | ${url.frequency}/${totalKeywords} (${coveragePct}%) |`);\n    }\n    lines.push('');\n  }\n\n  // Other tier \u2014 with keyword attribution\n  if (tiers.other.length > 0) {\n    lines.push(`### Other Results (${tiers.other.length})`);\n    lines.push('| # | URL | Score | Keywords |');\n    lines.push('|---|-----|-------|----------|');\n    for (const url of tiers.other) {\n      const keywords = url.queries.map(q => `\"${q}\"`).join(', ');\n      let domain: string;\n      try {\n        domain = new URL(url.url).hostname.replace(/^www\\./, '');\n      } catch {\n        domain = url.url;\n      }\n      lines.push(`| ${url.rank} | ${domain} | ${url.score.toFixed(1)} | ${keywords} |`);\n    }\n    lines.push('');\n  }\n\n  return lines.join('\\n');\n}\n\n// --- Metadata builder ---\n\nfunction buildMetadata(\n  aggregation: SearchAggregation,\n  executionTime: number,\n  totalKeywords: number,\n  searches: SearchResponse['searches'],\n  llmClassified: boolean,\n  llmError?: string,\n) {\n  const coverageSummary = searches.map(s => {\n    let topDomain: string | undefined;\n    const topResult = s.results[0];\n    if (topResult) {\n      try { topDomain = new URL(topResult.link).hostname.replace(/^www\\./, ''); } catch { /* ignore */ }\n    }\n    return { keyword: s.keyword, result_count: s.results.length, top_url: topDomain };\n  });\n  const lowYieldKeywords = searches\n    .filter(s => s.results.length <= 1)\n    .map(s => s.keyword);\n\n  return {\n    total_keywords: totalKeywords,\n    total_results: aggregation.rankedUrls.length,\n    execution_time_ms: executionTime,\n    total_unique_urls: aggregation.totalUniqueUrls,\n    consensus_url_count: aggregation.rankedUrls.filter(u => u.isConsensus).length,\n    frequency_threshold: aggregation.frequencyThreshold,\n    llm_classified: llmClassified,\n    ...(llmError ? { llm_error: llmError } : {}),\n    coverage_summary: coverageSummary,\n    ...(lowYieldKeywords.length > 0 ? { low_yield_keywords: lowYieldKeywords } : {}),\n  };\n}\n\n// --- Error builder ---\n\nfunction buildWebSearchError(\n  error: unknown,\n  params: WebSearchParams,\n  startTime: number,\n): ToolExecutionResult<WebSearchOutput> {\n  const structuredError = classifyError(error);\n  const executionTime = Date.now() - startTime;\n\n  mcpLog('error', `web-search: ${structuredError.message}`, 'search');\n\n  const errorContent = formatError({\n    code: structuredError.code,\n    message: structuredError.message,\n    retryable: structuredError.retryable,\n    toolName: 'web-search',\n    howToFix: ['Verify SERPER_API_KEY is set correctly'],\n    alternatives: [\n      'search-reddit(queries=[\"topic recommendations\"]) \u2014 returns Reddit URLs via Google search',\n      'scrape-links(urls=[...], what_to_extract=\"...\") \u2014 if you have URLs from prior steps, scrape them now',\n    ],\n  });\n\n  return toolFailure(\n    `${errorContent}\\n\\nExecution time: ${formatDuration(executionTime)}\\nKeywords: ${params.keywords.length}`,\n  );\n}\n\n// --- Main handler ---\n\nexport async function handleWebSearch(\n  params: WebSearchParams,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<WebSearchOutput>> {\n  const startTime = Date.now();\n\n  try {\n    mcpLog('info', `Searching for ${params.keywords.length} keyword(s)`, 'search');\n    await reporter.log('info', `Searching for ${params.keywords.length} keyword(s)`);\n    await reporter.progress(15, 100, 'Submitting search queries');\n\n    const response = await executeSearches(params.keywords);\n    await reporter.progress(50, 100, 'Collected search results');\n\n    const { aggregation } = processResults(response);\n    await reporter.log(\n      'info',\n      `Collected ${aggregation.totalUniqueUrls} unique URLs across ${response.totalKeywords} queries`,\n    );\n\n    // Decide: raw output or LLM classification\n    const useRaw = params.raw;\n    const llmProcessor = useRaw ? null : createLLMProcessor();\n\n    let markdown: string;\n    let llmClassified = false;\n    let llmError: string | undefined;\n\n    if (useRaw || !llmProcessor) {\n      // Raw path: traditional unified ranked list\n      if (!useRaw && !llmProcessor) {\n        llmError = 'LLM unavailable (LLM_EXTRACTION_API_KEY not set). Falling back to raw output.';\n        mcpLog('warning', llmError, 'search');\n      }\n      markdown = buildRawOutput(params.keywords, aggregation, response.searches);\n      await reporter.progress(80, 100, 'Ranking search results');\n    } else {\n      // LLM classification path\n      await reporter.progress(65, 100, 'Classifying results by relevance');\n      const classification = await classifySearchResults(\n        aggregation.rankedUrls,\n        params.objective,\n        response.totalKeywords,\n        llmProcessor,\n      );\n\n      if (classification.result) {\n        markdown = buildClassifiedOutput(\n          classification.result, aggregation, params.objective, response.totalKeywords,\n        );\n        llmClassified = true;\n        await reporter.progress(85, 100, 'Formatted classified results');\n      } else {\n        // Classification failed \u2014 fall back to raw\n        llmError = classification.error ?? 'Unknown classification error';\n        mcpLog('warning', `Classification failed, falling back to raw: ${llmError}`, 'search');\n        markdown = buildRawOutput(params.keywords, aggregation, response.searches);\n        await reporter.progress(85, 100, 'Classification failed, using raw output');\n      }\n    }\n\n    const executionTime = Date.now() - startTime;\n    const metadata = buildMetadata(\n      aggregation, executionTime, response.totalKeywords, response.searches, llmClassified, llmError,\n    );\n\n    mcpLog('info', `Search completed: ${aggregation.rankedUrls.length} URLs, classified=${llmClassified}`, 'search');\n    await reporter.log('info', `Search completed with ${aggregation.rankedUrls.length} URLs (classified: ${llmClassified})`);\n\n    const footer = `\\n---\\n*${formatDuration(executionTime)} | ${aggregation.totalUniqueUrls} unique URLs | ${metadata.consensus_url_count} consensus | threshold \u2265${aggregation.frequencyThreshold}${llmClassified ? ' | LLM classified' : ''}*`;\n    const fullMarkdown = markdown + footer;\n\n    return toolSuccess(fullMarkdown, { content: fullMarkdown, metadata });\n  } catch (error) {\n    return buildWebSearchError(error, params, startTime);\n  }\n}\n\nexport function registerWebSearchTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'web-search',\n      title: 'Web Search',\n      description:\n        'Search Google with 1-100 keywords in parallel. Each keyword runs as a separate search; results are aggregated, deduplicated, and ranked. Provide an objective describing what you need \u2014 an LLM classifies every result into 3 tiers (highly relevant / maybe relevant / other) based on titles and snippets alone (no URL fetching). Output: synthesis paragraph + tiered markdown table. Set raw=true to skip classification and get the traditional consensus-ranked URL list.',\n      schema: webSearchParamsSchema,\n      outputSchema: webSearchOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async (args, ctx) => {\n      if (!getCapabilities().search) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('search')));\n      }\n\n      const reporter = createToolReporter(ctx, 'web-search');\n      const result = await handleWebSearch(args, reporter);\n\n      await reporter.progress(100, 100, result.isError ? 'Search failed' : 'Search complete');\n      return toToolResponse(result);\n    },\n  );\n}\n"],
-  "mappings": "AAOA,SAAS,iBAAiB,4BAA4B;AACtD;AAAA,EACE;AAAA,EACA;AAAA,OAGK;AACP,SAAS,oBAAoB;AAC7B;AAAA,EACE;AAAA,EACA;AAAA,OACK;AACP,SAAS,oBAAoB,6BAAwD;AACrF,SAAS,qBAAqB;AAC9B;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,OACK;AACP;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AAkBP,eAAe,gBAAgB,UAA6C;AAC1E,QAAM,SAAS,IAAI,aAAa;AAChC,SAAO,OAAO,eAAe,QAAQ;AACvC;AAEA,SAAS,eAAe,UAGtB;AACA,QAAM,cAAc,iBAAiB,SAAS,UAAU,CAAC;AACzD,QAAM,gBAAgB,YAAY,WAAW,OAAO,OAAK,EAAE,WAAW;AACtE,SAAO,EAAE,aAAa,cAAc;AACtC;AAIA,SAAS,eACP,UACA,aACA,UACQ;AACR,SAAO;AAAA,IACL,YAAY;AAAA,IAAY;AAAA,IAAU;AAAA,IAClC,YAAY;AAAA,IACZ,YAAY;AAAA,IAAoB,YAAY;AAAA,EAC9C;AACF;AAIA,SAAS,sBACP,gBACA,aACA,WACA,eACQ;AACR,QAAM,aAAa,YAAY;AAG/B,QAAM,YAAY,IAAI,IAAI,WAAW,IAAI,OAAK,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC;AAG1D,QAAM,QAAQ;AAAA,IACZ,MAAM,CAAC;AAAA,IACP,OAAO,CAAC;AAAA,IACR,OAAO,CAAC;AAAA,EACV;AAGA,QAAM,UAAU,IAAI,IAAI,eAAe,QAAQ,IAAI,OAAK,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC;AAEzE,aAAW,OAAO,YAAY;AAC5B,UAAM,OAAO,QAAQ,IAAI,IAAI,IAAI;AACjC,QAAI,SAAS,mBAAmB;AAC9B,YAAM,KAAK,KAAK,GAAG;AAAA,IACrB,WAAW,SAAS,kBAAkB;AACpC,YAAM,MAAM,KAAK,GAAG;AAAA,IACtB,OAAO;AACL,YAAM,MAAM,KAAK,GAAG;AAAA,IACtB;AAAA,EACF;AAEA,QAAM,QAAkB,CAAC;AAGzB,QAAM,KAAK,MAAM,eAAe,KAAK,EAAE;AACvC,QAAM,KAAK,gBAAgB,SAAS,EAAE;AACtC,QAAM,KAAK,KAAK,aAAa,mBAAc,WAAW,MAAM,gBAAW,MAAM,KAAK,MAAM,qBAAqB,MAAM,MAAM,MAAM,oBAAoB;AACnJ,QAAM,KAAK,EAAE;AACb,QAAM,KAAK,gBAAgB,eAAe,SAAS,EAAE;AACrD,QAAM,KAAK,EAAE;AAGb,MAAI,MAAM,KAAK,SAAS,GAAG;AACzB,UAAM,KAAK,wBAAwB,MAAM,KAAK,MAAM,GAAG;AACvD,UAAM,KAAK,uBAAuB;AAClC,UAAM,KAAK,uBAAuB;AAClC,eAAW,OAAO,MAAM,MAAM;AAC5B,YAAM,cAAc,KAAK,MAAM,IAAI,gBAAgB,GAAG;AACtD,YAAM,UAAU,IAAI,QAAQ,IAAI,OAAK,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI;AACxD,YAAM,KAAK,KAAK,IAAI,IAAI,OAAO,IAAI,KAAK,KAAK,IAAI,GAAG,OAAO,IAAI,SAAS,IAAI,aAAa,KAAK,WAAW,MAAM;AAAA,IACjH;AACA,UAAM,KAAK,EAAE;AAAA,EACf;AAGA,MAAI,MAAM,MAAM,SAAS,GAAG;AAC1B,UAAM,KAAK,uBAAuB,MAAM,MAAM,MAAM,GAAG;AACvD,UAAM,KAAK,uBAAuB;AAClC,UAAM,KAAK,uBAAuB;AAClC,eAAW,OAAO,MAAM,OAAO;AAC7B,YAAM,cAAc,KAAK,MAAM,IAAI,gBAAgB,GAAG;AACtD,YAAM,KAAK,KAAK,IAAI,IAAI,OAAO,IAAI,KAAK,KAAK,IAAI,GAAG,OAAO,IAAI,SAAS,IAAI,aAAa,KAAK,WAAW,MAAM;AAAA,IACjH;AACA,UAAM,KAAK,EAAE;AAAA,EACf;AAGA,MAAI,MAAM,MAAM,SAAS,GAAG;AAC1B,UAAM,KAAK,sBAAsB,MAAM,MAAM,MAAM,GAAG;AACtD,UAAM,KAAK,gCAAgC;AAC3C,UAAM,KAAK,gCAAgC;AAC3C,eAAW,OAAO,MAAM,OAAO;AAC7B,YAAM,WAAW,IAAI,QAAQ,IAAI,OAAK,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI;AACzD,UAAI;AACJ,UAAI;AACF,iBAAS,IAAI,IAAI,IAAI,GAAG,EAAE,SAAS,QAAQ,UAAU,EAAE;AAAA,MACzD,QAAQ;AACN,iBAAS,IAAI;AAAA,MACf;AACA,YAAM,KAAK,KAAK,IAAI,IAAI,MAAM,MAAM,MAAM,IAAI,MAAM,QAAQ,CAAC,CAAC,MAAM,QAAQ,IAAI;AAAA,IAClF;AACA,UAAM,KAAK,EAAE;AAAA,EACf;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;AAIA,SAAS,cACP,aACA,eACA,eACA,UACA,eACA,UACA;AACA,QAAM,kBAAkB,SAAS,IAAI,OAAK;AACxC,QAAI;AACJ,UAAM,YAAY,EAAE,QAAQ,CAAC;AAC7B,QAAI,WAAW;AACb,UAAI;AAAE,oBAAY,IAAI,IAAI,UAAU,IAAI,EAAE,SAAS,QAAQ,UAAU,EAAE;AAAA,MAAG,QAAQ;AAAA,MAAe;AAAA,IACnG;AACA,WAAO,EAAE,SAAS,EAAE,SAAS,cAAc,EAAE,QAAQ,QAAQ,SAAS,UAAU;AAAA,EAClF,CAAC;AACD,QAAM,mBAAmB,SACtB,OAAO,OAAK,EAAE,QAAQ,UAAU,CAAC,EACjC,IAAI,OAAK,EAAE,OAAO;AAErB,SAAO;AAAA,IACL,gBAAgB;AAAA,IAChB,eAAe,YAAY,WAAW;AAAA,IACtC,mBAAmB;AAAA,IACnB,mBAAmB,YAAY;AAAA,IAC/B,qBAAqB,YAAY,WAAW,OAAO,OAAK,EAAE,WAAW,EAAE;AAAA,IACvE,qBAAqB,YAAY;AAAA,IACjC,gBAAgB;AAAA,IAChB,GAAI,WAAW,EAAE,WAAW,SAAS,IAAI,CAAC;AAAA,IAC1C,kBAAkB;AAAA,IAClB,GAAI,iBAAiB,SAAS,IAAI,EAAE,oBAAoB,iBAAiB,IAAI,CAAC;AAAA,EAChF;AACF;AAIA,SAAS,oBACP,OACA,QACA,WACsC;AACtC,QAAM,kBAAkB,cAAc,KAAK;AAC3C,QAAM,gBAAgB,KAAK,IAAI,IAAI;AAEnC,SAAO,SAAS,eAAe,gBAAgB,OAAO,IAAI,QAAQ;AAElE,QAAM,eAAe,YAAY;AAAA,IAC/B,MAAM,gBAAgB;AAAA,IACtB,SAAS,gBAAgB;AAAA,IACzB,WAAW,gBAAgB;AAAA,IAC3B,UAAU;AAAA,IACV,UAAU,CAAC,wCAAwC;AAAA,IACnD,cAAc;AAAA,MACZ;AAAA,MACA;AAAA,IACF;AAAA,EACF,CAAC;AAED,SAAO;AAAA,IACL,GAAG,YAAY;AAAA;AAAA,kBAAuB,eAAe,aAAa,CAAC;AAAA,YAAe,OAAO,SAAS,MAAM;AAAA,EAC1G;AACF;AAIA,eAAsB,gBACpB,QACA,WAAyB,eACsB;AAC/C,QAAM,YAAY,KAAK,IAAI;AAE3B,MAAI;AACF,WAAO,QAAQ,iBAAiB,OAAO,SAAS,MAAM,eAAe,QAAQ;AAC7E,UAAM,SAAS,IAAI,QAAQ,iBAAiB,OAAO,SAAS,MAAM,aAAa;AAC/E,UAAM,SAAS,SAAS,IAAI,KAAK,2BAA2B;AAE5D,UAAM,WAAW,MAAM,gBAAgB,OAAO,QAAQ;AACtD,UAAM,SAAS,SAAS,IAAI,KAAK,0BAA0B;AAE3D,UAAM,EAAE,YAAY,IAAI,eAAe,QAAQ;AAC/C,UAAM,SAAS;AAAA,MACb;AAAA,MACA,aAAa,YAAY,eAAe,uBAAuB,SAAS,aAAa;AAAA,IACvF;AAGA,UAAM,SAAS,OAAO;AACtB,UAAM,eAAe,SAAS,OAAO,mBAAmB;AAExD,QAAI;AACJ,QAAI,gBAAgB;AACpB,QAAI;AAEJ,QAAI,UAAU,CAAC,cAAc;AAE3B,UAAI,CAAC,UAAU,CAAC,cAAc;AAC5B,mBAAW;AACX,eAAO,WAAW,UAAU,QAAQ;AAAA,MACtC;AACA,iBAAW,eAAe,OAAO,UAAU,aAAa,SAAS,QAAQ;AACzE,YAAM,SAAS,SAAS,IAAI,KAAK,wBAAwB;AAAA,IAC3D,OAAO;AAEL,YAAM,SAAS,SAAS,IAAI,KAAK,kCAAkC;AACnE,YAAM,iBAAiB,MAAM;AAAA,QAC3B,YAAY;AAAA,QACZ,OAAO;AAAA,QACP,SAAS;AAAA,QACT;AAAA,MACF;AAEA,UAAI,eAAe,QAAQ;AACzB,mBAAW;AAAA,UACT,eAAe;AAAA,UAAQ;AAAA,UAAa,OAAO;AAAA,UAAW,SAAS;AAAA,QACjE;AACA,wBAAgB;AAChB,cAAM,SAAS,SAAS,IAAI,KAAK,8BAA8B;AAAA,MACjE,OAAO;AAEL,mBAAW,eAAe,SAAS;AACnC,eAAO,WAAW,+CAA+C,QAAQ,IAAI,QAAQ;AACrF,mBAAW,eAAe,OAAO,UAAU,aAAa,SAAS,QAAQ;AACzE,cAAM,SAAS,SAAS,IAAI,KAAK,yCAAyC;AAAA,MAC5E;AAAA,IACF;AAEA,UAAM,gBAAgB,KAAK,IAAI,IAAI;AACnC,UAAM,WAAW;AAAA,MACf;AAAA,MAAa;AAAA,MAAe,SAAS;AAAA,MAAe,SAAS;AAAA,MAAU;AAAA,MAAe;AAAA,IACxF;AAEA,WAAO,QAAQ,qBAAqB,YAAY,WAAW,MAAM,qBAAqB,aAAa,IAAI,QAAQ;AAC/G,UAAM,SAAS,IAAI,QAAQ,yBAAyB,YAAY,WAAW,MAAM,sBAAsB,aAAa,GAAG;AAEvH,UAAM,SAAS;AAAA;AAAA,GAAW,eAAe,aAAa,CAAC,MAAM,YAAY,eAAe,kBAAkB,SAAS,mBAAmB,gCAA2B,YAAY,kBAAkB,GAAG,gBAAgB,sBAAsB,EAAE;AAC1O,UAAM,eAAe,WAAW;AAEhC,WAAO,YAAY,cAAc,EAAE,SAAS,cAAc,SAAS,CAAC;AAAA,EACtE,SAAS,OAAO;AACd,WAAO,oBAAoB,OAAO,QAAQ,SAAS;AAAA,EACrD;AACF;AAEO,SAAS,sBAAsB,QAAyB;AAC7D,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,MAAM,QAAQ;AACnB,UAAI,CAAC,gBAAgB,EAAE,QAAQ;AAC7B,eAAO,eAAe,YAAY,qBAAqB,QAAQ,CAAC,CAAC;AAAA,MACnE;AAEA,YAAM,WAAW,mBAAmB,KAAK,YAAY;AACrD,YAAM,SAAS,MAAM,gBAAgB,MAAM,QAAQ;AAEnD,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,kBAAkB,iBAAiB;AACtF,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;",
+  "sourcesContent": ["/**\n * Web Search Tool Handler\n * NEVER throws - always returns structured response for graceful degradation\n */\n\nimport type { MCPServer } from 'mcp-use/server';\n\nimport { getCapabilities, getMissingEnvMessage } from '../config/index.js';\nimport {\n  webSearchOutputSchema,\n  webSearchParamsSchema,\n  type WebSearchParams,\n  type WebSearchOutput,\n} from '../schemas/web-search.js';\nimport { SearchClient } from '../clients/search.js';\nimport {\n  aggregateAndRank,\n  generateUnifiedOutput,\n} from '../utils/url-aggregator.js';\nimport { createLLMProcessor, classifySearchResults, type ClassificationResult } from '../services/llm-processor.js';\nimport { classifyError } from '../utils/errors.js';\nimport {\n  mcpLog,\n  formatError,\n  formatDuration,\n} from './utils.js';\nimport {\n  createToolReporter,\n  NOOP_REPORTER,\n  toolFailure,\n  toolSuccess,\n  toToolResponse,\n  type ToolExecutionResult,\n  type ToolReporter,\n} from './mcp-helpers.js';\n\n// --- Internal types ---\n\ninterface SearchAggregation {\n  readonly rankedUrls: ReturnType<typeof aggregateAndRank>['rankedUrls'];\n  readonly totalUniqueUrls: number;\n  readonly frequencyThreshold: number;\n  readonly thresholdNote?: string;\n}\n\ninterface SearchResponse {\n  searches: Parameters<typeof aggregateAndRank>[0];\n  totalKeywords: number;\n}\n\n// --- Helpers ---\n\nasync function executeSearches(keywords: string[]): Promise<SearchResponse> {\n  const client = new SearchClient();\n  return client.searchMultiple(keywords);\n}\n\nfunction processResults(response: SearchResponse): {\n  aggregation: SearchAggregation;\n  consensusUrls: SearchAggregation['rankedUrls'];\n} {\n  const aggregation = aggregateAndRank(response.searches, 5);\n  const consensusUrls = aggregation.rankedUrls.filter(u => u.isConsensus);\n  return { aggregation, consensusUrls };\n}\n\n// --- Raw output (traditional unified ranked list) ---\n\nfunction buildRawOutput(\n  keywords: string[],\n  aggregation: SearchAggregation,\n  searches: SearchResponse['searches'],\n): string {\n  return generateUnifiedOutput(\n    aggregation.rankedUrls, keywords, searches,\n    aggregation.totalUniqueUrls,\n    aggregation.frequencyThreshold, aggregation.thresholdNote,\n  );\n}\n\n// --- Classified output (3-tier LLM-classified table) ---\n\nfunction buildClassifiedOutput(\n  classification: ClassificationResult,\n  aggregation: SearchAggregation,\n  extract: string,\n  totalKeywords: number,\n): string {\n  const rankedUrls = aggregation.rankedUrls;\n\n  // Build lookup from rank \u2192 url data\n  const urlByRank = new Map(rankedUrls.map(u => [u.rank, u]));\n\n  // Build tier \u2192 entries mapping\n  const tiers = {\n    high: [] as typeof rankedUrls,\n    maybe: [] as typeof rankedUrls,\n    other: [] as typeof rankedUrls,\n  };\n\n  // Classify based on LLM response\n  const tierMap = new Map(classification.results.map(r => [r.rank, r.tier]));\n\n  for (const url of rankedUrls) {\n    const tier = tierMap.get(url.rank);\n    if (tier === 'HIGHLY_RELEVANT') {\n      tiers.high.push(url);\n    } else if (tier === 'MAYBE_RELEVANT') {\n      tiers.maybe.push(url);\n    } else {\n      tiers.other.push(url);\n    }\n  }\n\n  const lines: string[] = [];\n\n  // Header with generated title and synthesis\n  lines.push(`## ${classification.title}`);\n  lines.push(`> Looking for: ${extract}`);\n  lines.push(`> ${totalKeywords} queries \u2192 ${rankedUrls.length} URLs \u2192 ${tiers.high.length} highly relevant, ${tiers.maybe.length} possibly relevant`);\n  lines.push('');\n  lines.push(`**Summary:** ${classification.synthesis}`);\n  lines.push('');\n\n  // Highly Relevant tier\n  if (tiers.high.length > 0) {\n    lines.push(`### Highly Relevant (${tiers.high.length})`);\n    lines.push('| # | URL | Seen in |');\n    lines.push('|---|-----|---------|');\n    for (const url of tiers.high) {\n      const coveragePct = Math.round(url.coverageRatio * 100);\n      const queries = url.queries.map(q => `\"${q}\"`).join(', ');\n      lines.push(`| ${url.rank} | [${url.title}](${url.url}) | ${url.frequency}/${totalKeywords} (${coveragePct}%) |`);\n    }\n    lines.push('');\n  }\n\n  // Maybe Relevant tier\n  if (tiers.maybe.length > 0) {\n    lines.push(`### Maybe Relevant (${tiers.maybe.length})`);\n    lines.push('| # | URL | Seen in |');\n    lines.push('|---|-----|---------|');\n    for (const url of tiers.maybe) {\n      const coveragePct = Math.round(url.coverageRatio * 100);\n      lines.push(`| ${url.rank} | [${url.title}](${url.url}) | ${url.frequency}/${totalKeywords} (${coveragePct}%) |`);\n    }\n    lines.push('');\n  }\n\n  // Other tier \u2014 with keyword attribution\n  if (tiers.other.length > 0) {\n    lines.push(`### Other Results (${tiers.other.length})`);\n    lines.push('| # | URL | Score | Keywords |');\n    lines.push('|---|-----|-------|----------|');\n    for (const url of tiers.other) {\n      const keywords = url.queries.map(q => `\"${q}\"`).join(', ');\n      let domain: string;\n      try {\n        domain = new URL(url.url).hostname.replace(/^www\\./, '');\n      } catch {\n        domain = url.url;\n      }\n      lines.push(`| ${url.rank} | ${domain} | ${url.score.toFixed(1)} | ${keywords} |`);\n    }\n    lines.push('');\n  }\n\n  return lines.join('\\n');\n}\n\n// --- Metadata builder ---\n\nfunction buildMetadata(\n  aggregation: SearchAggregation,\n  executionTime: number,\n  totalKeywords: number,\n  searches: SearchResponse['searches'],\n  llmClassified: boolean,\n  llmError?: string,\n) {\n  const coverageSummary = searches.map(s => {\n    let topDomain: string | undefined;\n    const topResult = s.results[0];\n    if (topResult) {\n      try { topDomain = new URL(topResult.link).hostname.replace(/^www\\./, ''); } catch { /* ignore */ }\n    }\n    return { keyword: s.keyword, result_count: s.results.length, top_url: topDomain };\n  });\n  const lowYieldKeywords = searches\n    .filter(s => s.results.length <= 1)\n    .map(s => s.keyword);\n\n  return {\n    total_items: totalKeywords,\n    successful: aggregation.rankedUrls.length,\n    failed: totalKeywords - searches.filter(s => s.results.length > 0).length,\n    execution_time_ms: executionTime,\n    llm_classified: llmClassified,\n    ...(llmError ? { llm_error: llmError } : {}),\n    coverage_summary: coverageSummary,\n    ...(lowYieldKeywords.length > 0 ? { low_yield_keywords: lowYieldKeywords } : {}),\n  };\n}\n\n// --- Error builder ---\n\nfunction buildWebSearchError(\n  error: unknown,\n  params: WebSearchParams,\n  startTime: number,\n): ToolExecutionResult<WebSearchOutput> {\n  const structuredError = classifyError(error);\n  const executionTime = Date.now() - startTime;\n\n  mcpLog('error', `web-search: ${structuredError.message}`, 'search');\n\n  const errorContent = formatError({\n    code: structuredError.code,\n    message: structuredError.message,\n    retryable: structuredError.retryable,\n    toolName: 'web-search',\n    howToFix: ['Verify SERPER_API_KEY is set correctly'],\n    alternatives: [\n      'search-reddit(queries=[\"topic recommendations\"]) \u2014 returns Reddit URLs via Google search',\n      'scrape-links(urls=[...], what_to_extract=\"...\") \u2014 if you have URLs from prior steps, scrape them now',\n    ],\n  });\n\n  return toolFailure(\n    `${errorContent}\\n\\nExecution time: ${formatDuration(executionTime)}\\nQueries: ${params.queries.length}`,\n  );\n}\n\n// --- Main handler ---\n\nexport async function handleWebSearch(\n  params: WebSearchParams,\n  reporter: ToolReporter = NOOP_REPORTER,\n): Promise<ToolExecutionResult<WebSearchOutput>> {\n  const startTime = Date.now();\n\n  try {\n    mcpLog('info', `Searching for ${params.queries.length} query/queries`, 'search');\n    await reporter.log('info', `Searching for ${params.queries.length} query/queries`);\n    await reporter.progress(15, 100, 'Submitting search queries');\n\n    const response = await executeSearches(params.queries);\n    await reporter.progress(50, 100, 'Collected search results');\n\n    const { aggregation } = processResults(response);\n    await reporter.log(\n      'info',\n      `Collected ${aggregation.totalUniqueUrls} unique URLs across ${response.totalKeywords} queries`,\n    );\n\n    // Decide: raw output or LLM classification\n    const useRaw = params.raw;\n    const llmProcessor = useRaw ? null : createLLMProcessor();\n\n    let markdown: string;\n    let llmClassified = false;\n    let llmError: string | undefined;\n\n    if (useRaw || !llmProcessor) {\n      // Raw path: traditional unified ranked list\n      if (!useRaw && !llmProcessor) {\n        llmError = 'LLM unavailable (LLM_EXTRACTION_API_KEY not set). Falling back to raw output.';\n        mcpLog('warning', llmError, 'search');\n      }\n      markdown = buildRawOutput(params.queries, aggregation, response.searches);\n      await reporter.progress(80, 100, 'Ranking search results');\n    } else {\n      // LLM classification path\n      await reporter.progress(65, 100, 'Classifying results by relevance');\n      const classification = await classifySearchResults(\n        aggregation.rankedUrls,\n        params.extract,\n        response.totalKeywords,\n        llmProcessor,\n      );\n\n      if (classification.result) {\n        markdown = buildClassifiedOutput(\n          classification.result, aggregation, params.extract, response.totalKeywords,\n        );\n        llmClassified = true;\n        await reporter.progress(85, 100, 'Formatted classified results');\n      } else {\n        // Classification failed \u2014 fall back to raw\n        llmError = classification.error ?? 'Unknown classification error';\n        mcpLog('warning', `Classification failed, falling back to raw: ${llmError}`, 'search');\n        markdown = buildRawOutput(params.queries, aggregation, response.searches);\n        await reporter.progress(85, 100, 'Classification failed, using raw output');\n      }\n    }\n\n    const executionTime = Date.now() - startTime;\n    const metadata = buildMetadata(\n      aggregation, executionTime, response.totalKeywords, response.searches, llmClassified, llmError,\n    );\n\n    mcpLog('info', `Search completed: ${aggregation.rankedUrls.length} URLs, classified=${llmClassified}`, 'search');\n    await reporter.log('info', `Search completed with ${aggregation.rankedUrls.length} URLs (classified: ${llmClassified})`);\n\n    const footer = `\\n---\\n*${formatDuration(executionTime)} | ${aggregation.totalUniqueUrls} unique URLs${llmClassified ? ' | LLM classified' : ''}*`;\n    const fullMarkdown = markdown + footer;\n\n    return toolSuccess(fullMarkdown, { content: fullMarkdown, metadata });\n  } catch (error) {\n    return buildWebSearchError(error, params, startTime);\n  }\n}\n\nexport function registerWebSearchTool(server: MCPServer): void {\n  server.tool(\n    {\n      name: 'web-search',\n      title: 'Web Search',\n      description:\n        'Run up to 100 Google searches in parallel, aggregate and deduplicate results, then classify each URL by relevance to your extract goal. Returns a tiered table: highly relevant, maybe relevant, and other. Set raw=true for unclassified ranked results.',\n      schema: webSearchParamsSchema,\n      outputSchema: webSearchOutputSchema,\n      annotations: {\n        readOnlyHint: true,\n        idempotentHint: true,\n        destructiveHint: false,\n        openWorldHint: true,\n      },\n    },\n    async (args, ctx) => {\n      if (!getCapabilities().search) {\n        return toToolResponse(toolFailure(getMissingEnvMessage('search')));\n      }\n\n      const reporter = createToolReporter(ctx, 'web-search');\n      const result = await handleWebSearch(args, reporter);\n\n      await reporter.progress(100, 100, result.isError ? 'Search failed' : 'Search complete');\n      return toToolResponse(result);\n    },\n  );\n}\n"],
+  "mappings": "AAOA,SAAS,iBAAiB,4BAA4B;AACtD;AAAA,EACE;AAAA,EACA;AAAA,OAGK;AACP,SAAS,oBAAoB;AAC7B;AAAA,EACE;AAAA,EACA;AAAA,OACK;AACP,SAAS,oBAAoB,6BAAwD;AACrF,SAAS,qBAAqB;AAC9B;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,OACK;AACP;AAAA,EACE;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,EACA;AAAA,OAGK;AAkBP,eAAe,gBAAgB,UAA6C;AAC1E,QAAM,SAAS,IAAI,aAAa;AAChC,SAAO,OAAO,eAAe,QAAQ;AACvC;AAEA,SAAS,eAAe,UAGtB;AACA,QAAM,cAAc,iBAAiB,SAAS,UAAU,CAAC;AACzD,QAAM,gBAAgB,YAAY,WAAW,OAAO,OAAK,EAAE,WAAW;AACtE,SAAO,EAAE,aAAa,cAAc;AACtC;AAIA,SAAS,eACP,UACA,aACA,UACQ;AACR,SAAO;AAAA,IACL,YAAY;AAAA,IAAY;AAAA,IAAU;AAAA,IAClC,YAAY;AAAA,IACZ,YAAY;AAAA,IAAoB,YAAY;AAAA,EAC9C;AACF;AAIA,SAAS,sBACP,gBACA,aACA,SACA,eACQ;AACR,QAAM,aAAa,YAAY;AAG/B,QAAM,YAAY,IAAI,IAAI,WAAW,IAAI,OAAK,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC;AAG1D,QAAM,QAAQ;AAAA,IACZ,MAAM,CAAC;AAAA,IACP,OAAO,CAAC;AAAA,IACR,OAAO,CAAC;AAAA,EACV;AAGA,QAAM,UAAU,IAAI,IAAI,eAAe,QAAQ,IAAI,OAAK,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC;AAEzE,aAAW,OAAO,YAAY;AAC5B,UAAM,OAAO,QAAQ,IAAI,IAAI,IAAI;AACjC,QAAI,SAAS,mBAAmB;AAC9B,YAAM,KAAK,KAAK,GAAG;AAAA,IACrB,WAAW,SAAS,kBAAkB;AACpC,YAAM,MAAM,KAAK,GAAG;AAAA,IACtB,OAAO;AACL,YAAM,MAAM,KAAK,GAAG;AAAA,IACtB;AAAA,EACF;AAEA,QAAM,QAAkB,CAAC;AAGzB,QAAM,KAAK,MAAM,eAAe,KAAK,EAAE;AACvC,QAAM,KAAK,kBAAkB,OAAO,EAAE;AACtC,QAAM,KAAK,KAAK,aAAa,mBAAc,WAAW,MAAM,gBAAW,MAAM,KAAK,MAAM,qBAAqB,MAAM,MAAM,MAAM,oBAAoB;AACnJ,QAAM,KAAK,EAAE;AACb,QAAM,KAAK,gBAAgB,eAAe,SAAS,EAAE;AACrD,QAAM,KAAK,EAAE;AAGb,MAAI,MAAM,KAAK,SAAS,GAAG;AACzB,UAAM,KAAK,wBAAwB,MAAM,KAAK,MAAM,GAAG;AACvD,UAAM,KAAK,uBAAuB;AAClC,UAAM,KAAK,uBAAuB;AAClC,eAAW,OAAO,MAAM,MAAM;AAC5B,YAAM,cAAc,KAAK,MAAM,IAAI,gBAAgB,GAAG;AACtD,YAAM,UAAU,IAAI,QAAQ,IAAI,OAAK,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI;AACxD,YAAM,KAAK,KAAK,IAAI,IAAI,OAAO,IAAI,KAAK,KAAK,IAAI,GAAG,OAAO,IAAI,SAAS,IAAI,aAAa,KAAK,WAAW,MAAM;AAAA,IACjH;AACA,UAAM,KAAK,EAAE;AAAA,EACf;AAGA,MAAI,MAAM,MAAM,SAAS,GAAG;AAC1B,UAAM,KAAK,uBAAuB,MAAM,MAAM,MAAM,GAAG;AACvD,UAAM,KAAK,uBAAuB;AAClC,UAAM,KAAK,uBAAuB;AAClC,eAAW,OAAO,MAAM,OAAO;AAC7B,YAAM,cAAc,KAAK,MAAM,IAAI,gBAAgB,GAAG;AACtD,YAAM,KAAK,KAAK,IAAI,IAAI,OAAO,IAAI,KAAK,KAAK,IAAI,GAAG,OAAO,IAAI,SAAS,IAAI,aAAa,KAAK,WAAW,MAAM;AAAA,IACjH;AACA,UAAM,KAAK,EAAE;AAAA,EACf;AAGA,MAAI,MAAM,MAAM,SAAS,GAAG;AAC1B,UAAM,KAAK,sBAAsB,MAAM,MAAM,MAAM,GAAG;AACtD,UAAM,KAAK,gCAAgC;AAC3C,UAAM,KAAK,gCAAgC;AAC3C,eAAW,OAAO,MAAM,OAAO;AAC7B,YAAM,WAAW,IAAI,QAAQ,IAAI,OAAK,IAAI,CAAC,GAAG,EAAE,KAAK,IAAI;AACzD,UAAI;AACJ,UAAI;AACF,iBAAS,IAAI,IAAI,IAAI,GAAG,EAAE,SAAS,QAAQ,UAAU,EAAE;AAAA,MACzD,QAAQ;AACN,iBAAS,IAAI;AAAA,MACf;AACA,YAAM,KAAK,KAAK,IAAI,IAAI,MAAM,MAAM,MAAM,IAAI,MAAM,QAAQ,CAAC,CAAC,MAAM,QAAQ,IAAI;AAAA,IAClF;AACA,UAAM,KAAK,EAAE;AAAA,EACf;AAEA,SAAO,MAAM,KAAK,IAAI;AACxB;AAIA,SAAS,cACP,aACA,eACA,eACA,UACA,eACA,UACA;AACA,QAAM,kBAAkB,SAAS,IAAI,OAAK;AACxC,QAAI;AACJ,UAAM,YAAY,EAAE,QAAQ,CAAC;AAC7B,QAAI,WAAW;AACb,UAAI;AAAE,oBAAY,IAAI,IAAI,UAAU,IAAI,EAAE,SAAS,QAAQ,UAAU,EAAE;AAAA,MAAG,QAAQ;AAAA,MAAe;AAAA,IACnG;AACA,WAAO,EAAE,SAAS,EAAE,SAAS,cAAc,EAAE,QAAQ,QAAQ,SAAS,UAAU;AAAA,EAClF,CAAC;AACD,QAAM,mBAAmB,SACtB,OAAO,OAAK,EAAE,QAAQ,UAAU,CAAC,EACjC,IAAI,OAAK,EAAE,OAAO;AAErB,SAAO;AAAA,IACL,aAAa;AAAA,IACb,YAAY,YAAY,WAAW;AAAA,IACnC,QAAQ,gBAAgB,SAAS,OAAO,OAAK,EAAE,QAAQ,SAAS,CAAC,EAAE;AAAA,IACnE,mBAAmB;AAAA,IACnB,gBAAgB;AAAA,IAChB,GAAI,WAAW,EAAE,WAAW,SAAS,IAAI,CAAC;AAAA,IAC1C,kBAAkB;AAAA,IAClB,GAAI,iBAAiB,SAAS,IAAI,EAAE,oBAAoB,iBAAiB,IAAI,CAAC;AAAA,EAChF;AACF;AAIA,SAAS,oBACP,OACA,QACA,WACsC;AACtC,QAAM,kBAAkB,cAAc,KAAK;AAC3C,QAAM,gBAAgB,KAAK,IAAI,IAAI;AAEnC,SAAO,SAAS,eAAe,gBAAgB,OAAO,IAAI,QAAQ;AAElE,QAAM,eAAe,YAAY;AAAA,IAC/B,MAAM,gBAAgB;AAAA,IACtB,SAAS,gBAAgB;AAAA,IACzB,WAAW,gBAAgB;AAAA,IAC3B,UAAU;AAAA,IACV,UAAU,CAAC,wCAAwC;AAAA,IACnD,cAAc;AAAA,MACZ;AAAA,MACA;AAAA,IACF;AAAA,EACF,CAAC;AAED,SAAO;AAAA,IACL,GAAG,YAAY;AAAA;AAAA,kBAAuB,eAAe,aAAa,CAAC;AAAA,WAAc,OAAO,QAAQ,MAAM;AAAA,EACxG;AACF;AAIA,eAAsB,gBACpB,QACA,WAAyB,eACsB;AAC/C,QAAM,YAAY,KAAK,IAAI;AAE3B,MAAI;AACF,WAAO,QAAQ,iBAAiB,OAAO,QAAQ,MAAM,kBAAkB,QAAQ;AAC/E,UAAM,SAAS,IAAI,QAAQ,iBAAiB,OAAO,QAAQ,MAAM,gBAAgB;AACjF,UAAM,SAAS,SAAS,IAAI,KAAK,2BAA2B;AAE5D,UAAM,WAAW,MAAM,gBAAgB,OAAO,OAAO;AACrD,UAAM,SAAS,SAAS,IAAI,KAAK,0BAA0B;AAE3D,UAAM,EAAE,YAAY,IAAI,eAAe,QAAQ;AAC/C,UAAM,SAAS;AAAA,MACb;AAAA,MACA,aAAa,YAAY,eAAe,uBAAuB,SAAS,aAAa;AAAA,IACvF;AAGA,UAAM,SAAS,OAAO;AACtB,UAAM,eAAe,SAAS,OAAO,mBAAmB;AAExD,QAAI;AACJ,QAAI,gBAAgB;AACpB,QAAI;AAEJ,QAAI,UAAU,CAAC,cAAc;AAE3B,UAAI,CAAC,UAAU,CAAC,cAAc;AAC5B,mBAAW;AACX,eAAO,WAAW,UAAU,QAAQ;AAAA,MACtC;AACA,iBAAW,eAAe,OAAO,SAAS,aAAa,SAAS,QAAQ;AACxE,YAAM,SAAS,SAAS,IAAI,KAAK,wBAAwB;AAAA,IAC3D,OAAO;AAEL,YAAM,SAAS,SAAS,IAAI,KAAK,kCAAkC;AACnE,YAAM,iBAAiB,MAAM;AAAA,QAC3B,YAAY;AAAA,QACZ,OAAO;AAAA,QACP,SAAS;AAAA,QACT;AAAA,MACF;AAEA,UAAI,eAAe,QAAQ;AACzB,mBAAW;AAAA,UACT,eAAe;AAAA,UAAQ;AAAA,UAAa,OAAO;AAAA,UAAS,SAAS;AAAA,QAC/D;AACA,wBAAgB;AAChB,cAAM,SAAS,SAAS,IAAI,KAAK,8BAA8B;AAAA,MACjE,OAAO;AAEL,mBAAW,eAAe,SAAS;AACnC,eAAO,WAAW,+CAA+C,QAAQ,IAAI,QAAQ;AACrF,mBAAW,eAAe,OAAO,SAAS,aAAa,SAAS,QAAQ;AACxE,cAAM,SAAS,SAAS,IAAI,KAAK,yCAAyC;AAAA,MAC5E;AAAA,IACF;AAEA,UAAM,gBAAgB,KAAK,IAAI,IAAI;AACnC,UAAM,WAAW;AAAA,MACf;AAAA,MAAa;AAAA,MAAe,SAAS;AAAA,MAAe,SAAS;AAAA,MAAU;AAAA,MAAe;AAAA,IACxF;AAEA,WAAO,QAAQ,qBAAqB,YAAY,WAAW,MAAM,qBAAqB,aAAa,IAAI,QAAQ;AAC/G,UAAM,SAAS,IAAI,QAAQ,yBAAyB,YAAY,WAAW,MAAM,sBAAsB,aAAa,GAAG;AAEvH,UAAM,SAAS;AAAA;AAAA,GAAW,eAAe,aAAa,CAAC,MAAM,YAAY,eAAe,eAAe,gBAAgB,sBAAsB,EAAE;AAC/I,UAAM,eAAe,WAAW;AAEhC,WAAO,YAAY,cAAc,EAAE,SAAS,cAAc,SAAS,CAAC;AAAA,EACtE,SAAS,OAAO;AACd,WAAO,oBAAoB,OAAO,QAAQ,SAAS;AAAA,EACrD;AACF;AAEO,SAAS,sBAAsB,QAAyB;AAC7D,SAAO;AAAA,IACL;AAAA,MACE,MAAM;AAAA,MACN,OAAO;AAAA,MACP,aACE;AAAA,MACF,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,aAAa;AAAA,QACX,cAAc;AAAA,QACd,gBAAgB;AAAA,QAChB,iBAAiB;AAAA,QACjB,eAAe;AAAA,MACjB;AAAA,IACF;AAAA,IACA,OAAO,MAAM,QAAQ;AACnB,UAAI,CAAC,gBAAgB,EAAE,QAAQ;AAC7B,eAAO,eAAe,YAAY,qBAAqB,QAAQ,CAAC,CAAC;AAAA,MACnE;AAEA,YAAM,WAAW,mBAAmB,KAAK,YAAY;AACrD,YAAM,SAAS,MAAM,gBAAgB,MAAM,QAAQ;AAEnD,YAAM,SAAS,SAAS,KAAK,KAAK,OAAO,UAAU,kBAAkB,iBAAiB;AACtF,aAAO,eAAe,MAAM;AAAA,IAC9B;AAAA,EACF;AACF;",
   "names": []
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mcp-researchpowerpack-http",
-  "version": "4.0.0",
+  "version": "4.1.0",
   "description": "The ultimate research MCP toolkit: Reddit mining, web search with CTR aggregation, and intelligent web scraping - all in one modular package",
   "type": "module",
   "main": "dist/index.js",