mcp-scraper 0.24.0 → 0.25.0

package/README.md

@@ -88,7 +88,7 @@ Build the branded one-click bundle:
 npm run build:mcpb
 ```
 
-The generated bundle is written to `build/mcpb/mcp-scraper-<version>.mcpb` and copied to `public/downloads/` for the hosted download. The current public bundle is `https://mcpscraper.dev/downloads/mcp-scraper.mcpb` (`0.24.0`, SHA-256 `02f3b2a8ae2b264551fce9a2068eecd9cca22feed5946f077c32e84776138d1a`). Install it by opening or dragging it into Claude Desktop. Claude displays the `MCP Scraper` install card, icon, and API-key configuration field from the bundle manifest.
+The generated bundle is written to `build/mcpb/mcp-scraper-<version>.mcpb` and copied to `public/downloads/` for the hosted download. The current public bundle is `https://mcpscraper.dev/downloads/mcp-scraper.mcpb` (`0.25.0`, SHA-256 `357400ee624421d96671d7eef9fe050c35af691d1b42b160b86f5b86470679f1`). Install it by opening or dragging it into Claude Desktop. Claude displays the `MCP Scraper` install card, icon, and API-key configuration field from the bundle manifest.
 
 The MCPB install exposes every tool — web-intelligence plus all `browser_*` tools — through the one `mcp-scraper` server.
 
@@ -227,7 +227,7 @@ The `mcp-scraper` server (and the MCPB bundle, which runs it) exposes both secti
 
 All MCP tools expose output schemas and return `structuredContent` with the IDs, URLs, CSV paths, transcripts, browser session handles, replay paths, artifacts, recipe fields, or blueprint fields needed by the next step. Browser Agent tools keep a JSON text block for older clients, but structured data is the primary contract. All tools carry MCP annotations; file-writing tools such as replay downloads and annotations state their filesystem side effects.
 
-The canonical tool inventory is generated at `docs/mcp-tool-manifest.generated.json`. Both the `mcp-scraper` stdio server and the hosted endpoint at `https://mcpscraper.dev/mcp` expose the same 160 tools: 73 scraper, browser, workflow, billing, and connected-service tools plus 87 durable-memory tools. Release verification compares the exact local and remote tool-name sets, not only the count.
+The canonical tool inventory is generated at `docs/mcp-tool-manifest.generated.json`. Both the `mcp-scraper` stdio server and the hosted endpoint at `https://mcpscraper.dev/mcp` expose the same 162 tools: 75 scraper, browser, workflow, billing, and connected-service tools plus 87 durable-memory tools. Release verification compares the exact local and remote tool-name sets, not only the count.
 
 ## Resources
 

package/dist/bin/api-server.cjs

@@ -29200,7 +29200,7 @@ var PACKAGE_VERSION;
 var init_version = __esm({
   "src/version.ts"() {
     "use strict";
-    PACKAGE_VERSION = "0.24.0";
+    PACKAGE_VERSION = "0.25.0";
   }
 });
 
@@ -29315,7 +29315,7 @@ Multi-step orchestrations \u2014 prefer these over hand-chaining primitives when
 
 ## Memory
 mcp-scraper also exposes persistent per-user memory tools (notes, facts, vaults,
-scheduled actions, tables, channels) backed by memory.mcpscraper.dev. Every account starts with 13
+scheduled actions, tables, channels) backed by memory.mcpscraper.dev. Every account starts with 14
 vaults \u2014 call **list-vaults** to see what exists before creating anything new. Pick the vault whose job
 matches the content: **Ideas** (unvalidated concepts), **Knowledge** (distilled lessons/how-tos),
 **Library** (raw source material \u2014 articles, transcripts), **People** (one durable note per person/org),
@@ -29323,7 +29323,8 @@ matches the content: **Ideas** (unvalidated concepts), **Knowledge** (distilled
 (action items with a status), **Projects** (durable named initiatives), **Issues** (something broken),
 **Improvement Log** (receipts of changes made and whether they worked), **Experiments** (hypothesis +
 measured result), **Sprint** (current cycle's scope), **Inspiration** (other people's work worth
-studying).
+studying), **Skills** (reusable skill packages; each requires \`scripts/\` and at least one of
+\`references/\` or \`templates/\`, and may use both).
 For a normal new narrative note, call **prepare-memory-write** first. It queries the live vault contract,
 tag vocabulary, and natural neighbor vaults, and returns interlink candidates that must be read before
 acceptance. Compose every returned template section, then call **memory-capture**, which validates the
@@ -29353,6 +29354,12 @@ Tags are live vocabulary, not improvised labels: **list-memory-tags** shows what
 reusable, and has no exact, alias, or near-equivalent. Use **memory-backlinks**,
 **memory-graph-universe**, and **memory-graph-path** to trace the linked universe across vaults.
 
+**Always inspect the complete tag inventory and related notes first.** Use hybrid Smart RAG by default:
+form 3 focused queries (2\u20134 when useful), fuse exact tag/metadata/vault/date and semantic matches into 50
+candidates, expand the top 8 seeds by one link/backlink hop with at most 5 neighbors each, then Jina-rerank
+the combined pool to the best 30. Graph neighbors are candidates, never automatic links; add only links
+supported by the note contents.
+
 Scrape deposits are raw evidence and therefore go to **Library** through **library-ingest**, with the full
 Library template and source metadata. If the source contains durable applicable guidance, create a separate
 Knowledge companion through prepare-memory-write + memory-capture and link it to the Library source with
@@ -33794,7 +33801,7 @@ var init_browser_agent_mcp_server = __esm({
 });
 
 // src/mcp/memory-tool-schemas.ts
-var import_zod36, AcceptShareSchema, ApproveSenderSchema, DeclineShareSchema, GetChatLinkSchema, InboxSettingsSchema, InviteAccountSchema, issueKeyTool_scopeShape, IssueKeySchema, ListApprovedSendersSchema, ListKeysSchema, NoteInboxSchema, RemoveApprovedSenderSchema, RevokeChatLinkSchema, RevokeKeySchema, RevokeShareSchema, SetAgentIdentitySchema, SetScopeSchema, ShareNoteSchema, ShareVaultSchema, SwapVaultSchema, SwitchAccountSchema, UnlinkShareSchema, memoryCaptureTool_notePropsSchema, MemoryCaptureSchema, MemoryQuestionsSchema, PrepareMemoryWriteSchema, validateMemoryWriteTool_notePropsSchema, ValidateMemoryWriteSchema, CreateChannelSchema, GetMessageNoteSchema, ListChannelMembersSchema, ListChannelMessagesSchema, MyMentionsSchema, PollChannelSchema, PostMessageSchema, ReactMessageSchema, RemoveChannelMemberSchema, ReplyMessageSchema, factHistoryTool_entryShape, FactHistorySchema, recordFactTool_factShape, RecordFactSchema, NoteBacklinksSchema, GraphPathSchema, GraphUniverseSchema, LibraryIngestSchema, DeleteNoteSchema, ExportSchema, getTool_notePropsSchema, GetSchema, ListSchema, putTool_notePropsSchema, PutSchema, SearchSchema, SuggestSchema, UploadSchema, TemporalRecallSchema, CreateScheduledActionSchema, DeleteScheduledActionSchema, GetScheduleLinkSchema, GetScheduleStatusSchema, ListScheduledActionsSchema, PauseScheduledActionSchema, ProposeScheduledActionSchema, ResumeScheduledActionSchema, RevokeScheduleLinkSchema, SetScheduleEntitlementSchema, CostUsageSchema, StorageUsageSchema, CreateTableSchema, deleteTableRowsTool_FilterSchema, DeleteTableRowsSchema, DescribeTableSchema, DropTableSchema, InsertTableRowsSchema, ListTablesSchema, queryTableTool_FilterSchema, QueryTableSchema, ListTagsSchema, ResolveTagsSchema, UpsertTagSchema, AddVaultSchema, CreateSecureVaultSchema, DeleteVaultSchema, GetVaultContractSchema, ListSharedWithMeSchema, ListVaultsSchema, ProvisionDefaultsSchema, RouteMemorySchema, VideoAnalyzeStartSchema, VideoAnalyzeStatusSchema, CreateWebhookSchema, ListWebhooksSchema, RevokeWebhookSchema, MEMORY_TOOL_SCHEMAS;
+var import_zod36, AcceptShareSchema, ApproveSenderSchema, DeclineShareSchema, GetChatLinkSchema, InboxSettingsSchema, InviteAccountSchema, issueKeyTool_scopeShape, IssueKeySchema, ListApprovedSendersSchema, ListKeysSchema, NoteInboxSchema, RemoveApprovedSenderSchema, RevokeChatLinkSchema, RevokeKeySchema, RevokeShareSchema, SetAgentIdentitySchema, SetScopeSchema, ShareNoteSchema, ShareVaultSchema, SwapVaultSchema, SwitchAccountSchema, UnlinkShareSchema, memoryCaptureTool_notePropsSchema, MemoryCaptureSchema, MemoryQuestionsSchema, PrepareMemoryWriteSchema, validateMemoryWriteTool_notePropsSchema, ValidateMemoryWriteSchema, CreateChannelSchema, GetMessageNoteSchema, ListChannelMembersSchema, ListChannelMessagesSchema, MyMentionsSchema, PollChannelSchema, PostMessageSchema, ReactMessageSchema, RemoveChannelMemberSchema, ReplyMessageSchema, factHistoryTool_entryShape, FactHistorySchema, recordFactTool_factShape, RecordFactSchema, NoteBacklinksSchema, GraphPathSchema, GraphUniverseSchema, LibraryIngestSchema, bulkDeleteNotesTool_FilterSchema, BulkDeleteNotesSchema, DeleteNoteSchema, ExportSchema, getTool_notePropsSchema, GetSchema, ListSchema, putTool_notePropsSchema, PutSchema, searchTool_primitiveValue, SearchSchema, SuggestSchema, UploadSchema, TemporalRecallSchema, CreateScheduledActionSchema, DeleteScheduledActionSchema, GetScheduleLinkSchema, GetScheduleStatusSchema, ListScheduledActionsSchema, PauseScheduledActionSchema, ProposeScheduledActionSchema, ResumeScheduledActionSchema, RevokeScheduleLinkSchema, SetScheduleDefaultsSchema, SetScheduleEntitlementSchema, CostUsageSchema, StorageUsageSchema, CreateTableSchema, deleteTableRowsTool_FilterSchema, DeleteTableRowsSchema, DescribeTableSchema, DropTableSchema, InsertTableRowsSchema, ListTablesSchema, queryTableTool_FilterSchema, QueryTableSchema, ListTagsSchema, ResolveTagsSchema, UpsertTagSchema, AddVaultSchema, CreateSecureVaultSchema, DeleteVaultSchema, GetVaultContractSchema, ListSharedWithMeSchema, ListVaultsSchema, ProvisionDefaultsSchema, RouteMemorySchema, VideoAnalyzeStartSchema, VideoAnalyzeStatusSchema, CreateWebhookSchema, ListWebhooksSchema, RevokeWebhookSchema, MEMORY_TOOL_SCHEMAS;
 var init_memory_tool_schemas = __esm({
   "src/mcp/memory-tool-schemas.ts"() {
     "use strict";
@@ -34318,12 +34325,15 @@ var init_memory_tool_schemas = __esm({
       type: import_zod36.z.string().describe("Note type from the target vault contract (also used to route the note)."),
       domain: import_zod36.z.string().describe("Domain folder for Library/Knowledge (AI, SEO, Copywriting & Ads, Business, Spirituality)."),
       folder: import_zod36.z.string().describe("Explicit sub-folder within the vault; overrides routing-derived folder."),
+      script_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: one or more Obsidian/internal paths under scripts/."),
+      reference_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under references/."),
+      template_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under templates/."),
       parentMessageId: import_zod36.z.string().describe("Channel messages only: the path of the top-level message this is a reply to. Absent on top-level messages.")
     }).partial().passthrough();
     MemoryCaptureSchema = {
       id: "memory-capture",
       upstreamName: "memoryCaptureTool",
-      description: "Strict normal-create path for durable memory. Refuses incomplete notes, writes through memory-put, registers canonical tags, and reads the note back to verify persisted content and props. Call prepare-memory-write first. Reserve memory-put for low-level migrations or deliberate edits.",
+      description: "Strict normal-create path for durable memory. Before capture, list the complete tag vocabulary, call prepare-memory-write, run hybrid memory-search for related notes, and read the strongest link candidates. This tool refuses incomplete notes, writes through memory-put, registers canonical tags, and verifies persisted content and props. Reserve memory-put for low-level migrations or deliberate edits.",
       input: {
         vault: import_zod36.z.string(),
         folder: import_zod36.z.string().optional(),
@@ -34381,7 +34391,7 @@ var init_memory_tool_schemas = __esm({
     PrepareMemoryWriteSchema = {
       id: "prepare-memory-write",
       upstreamName: "prepareMemoryWriteTool",
-      description: "Mandatory planning pass for a normal new memory. Routes the note, returns the live template and natural vault relationships, resolves proposed tags against the registry, and shortlists interlink opportunities that must be read/reviewed before capture.",
+      description: "Mandatory planning pass for a normal new memory. First inspect the complete tag vocabulary with list-memory-tags; then this pass routes the note, returns the live template and natural vault relationships, resolves proposed tags, and shortlists interlinks. Use hybrid memory-search (3 focused queries, 50 fused candidates, bounded graph expansion, rerank to 30 by default) and read strong related notes before capture. This is an explicit AI workflow directive, not a claim of persisted call-order enforcement.",
       input: {
         title: import_zod36.z.string().min(1),
         content: import_zod36.z.string().min(1),
@@ -34417,6 +34427,9 @@ var init_memory_tool_schemas = __esm({
       type: import_zod36.z.string().describe("Note type from the target vault contract (also used to route the note)."),
       domain: import_zod36.z.string().describe("Domain folder for Library/Knowledge (AI, SEO, Copywriting & Ads, Business, Spirituality)."),
       folder: import_zod36.z.string().describe("Explicit sub-folder within the vault; overrides routing-derived folder."),
+      script_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: one or more Obsidian/internal paths under scripts/."),
+      reference_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under references/."),
+      template_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under templates/."),
       parentMessageId: import_zod36.z.string().describe("Channel messages only: the path of the top-level message this is a reply to. Absent on top-level messages.")
     }).partial().passthrough();
     ValidateMemoryWriteSchema = {
@@ -34872,6 +34885,39 @@ var init_memory_tool_schemas = __esm({
         openWorldHint: true
       }
     };
+    bulkDeleteNotesTool_FilterSchema = import_zod36.z.object({
+      column: import_zod36.z.enum(["path", "title", "kind", "source", "captured_at", "created_at", "updated_at", "revision"]).describe("Note column to filter on."),
+      op: import_zod36.z.enum(["eq", "neq", "gt", "gte", "lt", "lte", "like", "prefix", "in"]).describe('"prefix" anchors to the start of the value only (e.g. path prefix "Mastra/opt-"); "like" matches anywhere in the value; "in" requires an array value.'),
+      value: import_zod36.z.unknown().describe('Value to compare against. For "in", pass an array.')
+    });
+    BulkDeleteNotesSchema = {
+      id: "bulk-delete-notes",
+      upstreamName: "bulkDeleteNotesTool",
+      description: "Delete every note in one vault matching all given filters (ANDed) plus an optional tag match (ANY of the given tags) \u2014 at least one filter or tag is required; there is no unfiltered delete-all here (use delete-vault for that). Defaults to dryRun: true, previewing the match count and a sample of up to 20 matching notes without deleting anything. Pass dryRun: false to actually delete. DESTRUCTIVE and not recoverable when dryRun is false. Requires write scope. Refuses secure vaults (not indexed, filtering is meaningless there).",
+      input: {
+        vault: import_zod36.z.string().optional().describe("Vault to delete from. Optional; defaults to the session active vault, then the first vault the caller is entitled to."),
+        filters: import_zod36.z.array(bulkDeleteNotesTool_FilterSchema).default([]).describe("Filters to AND together. At least one of filters/tags is required."),
+        tags: import_zod36.z.array(import_zod36.z.string()).optional().describe("Match notes carrying ANY of these tags (ORed among themselves, ANDed with filters). At least one of filters/tags is required."),
+        dryRun: import_zod36.z.boolean().default(true).describe("When true (default), previews matches without deleting anything. Pass false to actually delete.")
+      },
+      output: {
+        ok: import_zod36.z.boolean().describe("True on success; false on auth/scope/validation error."),
+        vault: import_zod36.z.string().optional().describe("The vault operated on (after defaulting). Present when ok is true."),
+        dryRun: import_zod36.z.boolean().optional().describe("Echoes whether this call was a preview or a real delete."),
+        matchCount: import_zod36.z.number().optional().describe("Number of notes matching the filters. Present on a dry run."),
+        sample: import_zod36.z.array(import_zod36.z.object({ path: import_zod36.z.string(), title: import_zod36.z.string(), updatedAt: import_zod36.z.string() })).optional().describe("Up to 20 matching notes, most-recently-updated first. Present on a dry run."),
+        deleted: import_zod36.z.number().optional().describe("Number of notes actually deleted. Present when dryRun is false."),
+        deletedPaths: import_zod36.z.array(import_zod36.z.string()).optional().describe("Paths of the deleted notes. Present when dryRun is false."),
+        error: import_zod36.z.string().optional().describe("Human-readable failure reason when ok is false.")
+      },
+      annotations: {
+        title: "Bulk Delete Notes",
+        readOnlyHint: false,
+        destructiveHint: true,
+        idempotentHint: false,
+        openWorldHint: false
+      }
+    };
     DeleteNoteSchema = {
       id: "delete-note",
       upstreamName: "deleteNoteTool",
@@ -34953,6 +34999,9 @@ var init_memory_tool_schemas = __esm({
       type: import_zod36.z.string().describe("Note type from the target vault contract (also used to route the note)."),
       domain: import_zod36.z.string().describe("Domain folder for Library/Knowledge (AI, SEO, Copywriting & Ads, Business, Spirituality)."),
       folder: import_zod36.z.string().describe("Explicit sub-folder within the vault; overrides routing-derived folder."),
+      script_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: one or more Obsidian/internal paths under scripts/."),
+      reference_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under references/."),
+      template_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under templates/."),
       parentMessageId: import_zod36.z.string().describe("Channel messages only: the path of the top-level message this is a reply to. Absent on top-level messages.")
     }).partial().passthrough();
     GetSchema = {
@@ -35041,6 +35090,9 @@ var init_memory_tool_schemas = __esm({
       type: import_zod36.z.string().describe("Note type from the target vault contract (also used to route the note)."),
       domain: import_zod36.z.string().describe("Domain folder for Library/Knowledge (AI, SEO, Copywriting & Ads, Business, Spirituality)."),
       folder: import_zod36.z.string().describe("Explicit sub-folder within the vault; overrides routing-derived folder."),
+      script_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: one or more Obsidian/internal paths under scripts/."),
+      reference_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under references/."),
+      template_links: import_zod36.z.array(import_zod36.z.string()).describe("Skills only: optional Obsidian/internal paths under templates/."),
       parentMessageId: import_zod36.z.string().describe("Channel messages only: the path of the top-level message this is a reply to. Absent on top-level messages.")
     }).partial().passthrough();
     PutSchema = {
@@ -35085,44 +35137,64 @@ var init_memory_tool_schemas = __esm({
         openWorldHint: true
       }
     };
+    searchTool_primitiveValue = import_zod36.z.union([import_zod36.z.string(), import_zod36.z.number(), import_zod36.z.boolean()]);
     SearchSchema = {
       id: "memory-search",
       upstreamName: "searchTool",
-      description: "Semantic search across every vault the caller can reach (own, shared, channels) in one call \u2014 reach for this first when searching by meaning. For a fast title-only 'does this exist' check, use memory-suggest instead. Form query as a well-phrased semantic reformulation (expand abbreviations, add synonyms/entities) rather than the user's raw message; pass their original wording separately in userMessage for retrieval-quality logging. Each result is tagged with its source vault. Embeds the query (network call).",
+      description: "Default Smart RAG search across accessible memory. Form 2-4 focused query variants, combine semantic matches with exact vault/tag/date/kind/type/metadata filters, expand one bounded hop of outgoing links and backlinks around strong seeds, then rerank. Defaults: retrieve/fuse 50 candidates, 8 graph seeds, 5 neighbors per seed, rerank to 30. Graph neighbors are candidates, never automatic winners or links. Before tagging or writing, also call list-memory-tags to inspect the complete vocabulary and reuse existing tags; read strong related notes before selecting links.",
       input: {
-        vault: import_zod36.z.string().optional().describe(
-          "Narrow the search to exactly this vault. Optional; when omitted (the default and recommended usage) every vault the caller is entitled to is searched in one call."
-        ),
-        query: import_zod36.z.string().min(1).describe(
-          "The semantic search query \u2014 YOUR reformulation of what the human wants, not their raw message verbatim. Expand ambiguous references, add relevant synonyms/entities. Must be non-empty."
-        ),
-        userMessage: import_zod36.z.string().optional().describe("The human's original, unmodified message that prompted this search. Optional but recommended \u2014 recorded for retrieval-quality review, does not affect the search."),
-        topK: import_zod36.z.number().int().min(1).max(50).optional().describe("Maximum number of chunks to return. Optional; default 8, range 1-50."),
-        includeShared: import_zod36.z.boolean().optional().describe("Whether to also search notes individually shared with you and accepted. Default true.")
+        vault: import_zod36.z.string().optional().describe("Exact logical vault handle to search. Omit to search every entitled vault."),
+        query: import_zod36.z.string().min(1).describe("A focused semantic reformulation of the request."),
+        userMessage: import_zod36.z.string().optional().describe("Original human wording. Used as a distinct query variant when useful and logged for quality review."),
+        queries: import_zod36.z.array(import_zod36.z.string().min(1)).min(2).max(4).optional().describe("Caller-provided focused query variants. The planner deduplicates these with query, original wording, and tag/entity terms."),
+        entities: import_zod36.z.array(import_zod36.z.string().min(1)).max(20).optional().describe("Named entities or exact terms used to form a focused query variant."),
+        tags: import_zod36.z.array(import_zod36.z.string().min(1)).max(20).optional().describe("Exact canonical tags used for tag-channel retrieval and filtering."),
+        tagMode: import_zod36.z.enum(["any", "all"]).optional().describe("Whether a note must match any or all supplied tags. Default any."),
+        kind: import_zod36.z.enum(["note", "library", "capture", "decision", "message"]).optional().describe("Exact persisted note-kind filter."),
+        type: import_zod36.z.string().min(1).optional().describe("Exact props.type filter."),
+        dateFrom: import_zod36.z.string().datetime().optional().describe("Inclusive lower bound for note updatedAt."),
+        dateTo: import_zod36.z.string().datetime().optional().describe("Inclusive upper bound for note updatedAt."),
+        metadata: import_zod36.z.record(searchTool_primitiveValue).optional().describe("Exact equality filters for primitive note props."),
+        queryVariantCount: import_zod36.z.number().int().min(2).max(4).optional().describe("Number of variants. Default 3."),
+        candidatePool: import_zod36.z.number().int().min(10).max(100).optional().describe("Total fused candidates before final reranking. Default 50."),
+        graphSeedCount: import_zod36.z.number().int().min(0).max(20).optional().describe("Strong preliminary notes whose graph neighborhoods are considered. Default 8."),
+        graphDepth: import_zod36.z.literal(1).optional().describe("Graph expansion depth. Bounded to exactly one hop."),
+        graphNeighborsPerSeed: import_zod36.z.number().int().min(1).max(10).optional().describe("Maximum outgoing-link plus backlink neighbors per seed. Default 5."),
+        rerankTopN: import_zod36.z.number().int().min(1).max(50).optional().describe("Final results retained after Jina reranking. Default 30."),
+        topK: import_zod36.z.number().int().min(1).max(50).optional().describe("Deprecated compatibility alias for rerankTopN. Prefer rerankTopN; default remains 30."),
+        includeShared: import_zod36.z.boolean().optional().describe("Also search individually accepted shares. Default true. Exact note metadata filters exclude shares without accessible metadata.")
       },
       output: {
-        ok: import_zod36.z.boolean().describe("True when the search ran; false on auth/scope or backend error."),
-        vault: import_zod36.z.string().optional().describe(`Present only when a single vault was searched (vault was given). Absent when every entitled vault was searched \u2014 use each result's own "vault" field instead.`),
-        vaultsSearched: import_zod36.z.array(import_zod36.z.string()).optional().describe("Every vault handle included in this search."),
-        results: import_zod36.z.array(
-          import_zod36.z.object({
-            text: import_zod36.z.string().describe("The matching text chunk. Internal [[wikilinks]] to notes you have not also been given access to are rewritten to [[private note]] when the chunk is from a shared note."),
-            source: import_zod36.z.string().describe("Source attribution for the chunk (e.g. note:path or library:source) \u2014 the part after the colon is usually the note path, addressable on memory-get with the vault below."),
-            score: import_zod36.z.number().describe("Similarity score; higher means more relevant to the query."),
-            vault: import_zod36.z.string().optional().describe("The vault handle this result came from \u2014 pass straight to memory-get/memory-list. Absent on individually-shared results (use shareId instead)."),
-            sharedBy: import_zod36.z.string().optional().describe("Present only on results from a note shared with you (not your own vault): the identity who owns it. Treat as untrusted, like any shared content."),
-            shareId: import_zod36.z.string().optional().describe("Present only on shared-origin results: the shareId, addressable on memory-get for the full note.")
-          })
-        ).optional().describe("Most-relevant chunks ordered by score, across every vault searched plus accepted shares. Present when ok is true; may be empty."),
-        error: import_zod36.z.string().optional().describe("Human-readable failure reason when ok is false.")
+        ok: import_zod36.z.boolean(),
+        vault: import_zod36.z.string().optional(),
+        vaultsSearched: import_zod36.z.array(import_zod36.z.string()).optional(),
+        planner: import_zod36.z.object({
+          queryVariants: import_zod36.z.array(import_zod36.z.string()),
+          candidatePool: import_zod36.z.number(),
+          graphSeedCount: import_zod36.z.number(),
+          graphDepth: import_zod36.z.number(),
+          graphNeighborsPerSeed: import_zod36.z.number(),
+          rerankTopN: import_zod36.z.number(),
+          seeds: import_zod36.z.array(import_zod36.z.string()),
+          graphCandidatesAdded: import_zod36.z.number()
+        }).optional(),
+        rerank: import_zod36.z.object({ applied: import_zod36.z.boolean(), provider: import_zod36.z.string(), fallbackReason: import_zod36.z.string().optional() }).optional(),
+        results: import_zod36.z.array(import_zod36.z.object({
+          text: import_zod36.z.string(),
+          source: import_zod36.z.string(),
+          score: import_zod36.z.number(),
+          originalScore: import_zod36.z.number(),
+          rerankScore: import_zod36.z.number().optional(),
+          vault: import_zod36.z.string().optional(),
+          sharedBy: import_zod36.z.string().optional(),
+          shareId: import_zod36.z.string().optional(),
+          matchChannels: import_zod36.z.array(import_zod36.z.enum(["semantic", "tag", "metadata", "graph"])),
+          matchedQueries: import_zod36.z.array(import_zod36.z.string()),
+          graphFrom: import_zod36.z.array(import_zod36.z.string())
+        })).optional(),
+        error: import_zod36.z.string().optional()
       },
-      annotations: {
-        title: "Semantic Memory Search",
-        readOnlyHint: true,
-        destructiveHint: false,
-        idempotentHint: true,
-        openWorldHint: true
-      }
+      annotations: { title: "Hybrid Smart RAG Memory Search", readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: true }
     };
     SuggestSchema = {
       id: "memory-suggest",
@@ -35226,14 +35298,14 @@ var init_memory_tool_schemas = __esm({
     CreateScheduledActionSchema = {
       id: "create-scheduled-action",
       upstreamName: "createScheduledActionTool",
-      description: "Create a Credit-metered scheduled action for an active MCP Scraper Starter plan or higher, in agent mode (default) or connection_sync mode. Each execution has a 75-Credit base charge; agent model usage is added at 1.5 times OpenRouter's actual reported cost. Agent mode follows the description and writes a result into the target vault. connection_sync deterministically runs approved read-only tools on bound service connections and ingests their data. Google Search Console syncs also upsert a typed tenant-owned performance table for exact filtering with table-query; discover its tableName by calling list_service_connections after the first successful run. Cadence 'once' runs a single time then completes permanently. Requires write access to the target vault.",
+      description: "Create a Credit-metered scheduled action for an active MCP Scraper Starter plan or higher, in agent mode (default) or connection_sync mode. Each execution has a 75-Credit base charge; agent model usage is added at 1.5 times OpenRouter's actual reported cost. Agent mode follows the description and writes a result into the target vault. connection_sync deterministically runs the approved read-only tools on bound service connections and ingests their data; it requires at least one connection to be bound before execution. Google Search Console syncs also upsert a typed tenant-owned performance table for exact filtering with table-query; discover its tableName by calling list_service_connections after the first successful run. Cadence 'once' runs a single time then completes permanently. Requires write access to the target vault.",
       input: {
         description: import_zod36.z.string().min(1).describe("Free-text description of what this action should do each time it runs."),
         vault: import_zod36.z.string().min(1).describe("The vault this action writes its results into. You must already have write access to it."),
         cadence: import_zod36.z.enum(["once", "daily", "weekly", "monthly"]).describe('How often this action runs. "once" fires a single time and then completes.'),
         executionMode: import_zod36.z.enum(["agent", "connection_sync"]).default("agent").describe(`How to execute each run. "agent" (default) lets an agent follow the description. "connection_sync" deterministically ingests data from the schedule's bound connections using only their approved read-only tools; bind at least one connection before it runs. Search Console connection_sync also maintains a typed table exposed as the connection tableName.`),
         timeOfDay: import_zod36.z.string().regex(/^([01]\d|2[0-3]):([0-5]\d)$/).optional().describe("24-hour HH:MM clock time to run at, in the given timezone. Optional \u2014 omit to run at any time during the period (matches prior default behavior)."),
-        timezone: import_zod36.z.string().optional().describe('IANA timezone name, e.g. "America/Denver". Only meaningful together with timeOfDay. Defaults to UTC.'),
+        timezone: import_zod36.z.string().optional().describe(`IANA timezone name, e.g. "America/Denver". Only meaningful together with timeOfDay or deployDate. Omit to use the account's default timezone (set via set-schedule-defaults), falling back to UTC.`),
         deployDate: import_zod36.z.string().regex(/^\d{4}-\d{2}-\d{2}$/).optional().describe('Calendar date (YYYY-MM-DD, in the given timezone) this action should first become eligible to run \u2014 its deployment/start date. For recurring cadences, the first occurrence lands on or after this date; every later occurrence still follows the normal cadence. For cadence "once", this (combined with timeOfDay if given) is exactly what day it fires. Omit to start immediately.')
       },
       output: {
@@ -35295,7 +35367,7 @@ var init_memory_tool_schemas = __esm({
     GetScheduleStatusSchema = {
       id: "get-schedule-status",
       upstreamName: "getScheduleStatusTool",
-      description: "Get the Credit-metered Scheduled Actions access and billing policy. Scheduling requires an active MCP Scraper Starter plan or higher but has no separate subscription: each execution has a 75-Credit base charge, and agent model usage is billed at 1.5 times OpenRouter's actual reported cost.",
+      description: "Get the Credit-metered Scheduled Actions access, billing policy, and default timezone. Scheduling requires an active MCP Scraper Starter plan or higher but has no separate subscription: each execution has a 75-Credit base charge, and agent model usage is billed at 1.5 times OpenRouter's actual reported cost.",
       input: {},
       output: {
         ok: import_zod36.z.boolean(),
@@ -35371,7 +35443,7 @@ var init_memory_tool_schemas = __esm({
     ProposeScheduledActionSchema = {
       id: "propose-scheduled-action",
       upstreamName: "proposeScheduledActionTool",
-      description: "Turn freeform text describing what you want automated into a structured proposal (description, vault, cadence) for review and confirmation via create-scheduled-action \u2014 this is the propose step only; nothing is created here.",
+      description: "Turn freeform text describing what you want automated into a structured proposal (description, vault, cadence, time of day, timezone) for review and confirmation via create-scheduled-action \u2014 this is the propose step only; nothing is created here.",
       input: {
         request: import_zod36.z.string().min(1).describe("Freeform text describing what you want scheduled.")
       },
@@ -35380,6 +35452,8 @@ var init_memory_tool_schemas = __esm({
         description: import_zod36.z.string().optional().describe("Proposed description \u2014 edit freely before confirming."),
         vault: import_zod36.z.string().optional().describe("Proposed target vault."),
         cadence: import_zod36.z.string().optional().describe("Proposed cadence: once, daily, weekly, or monthly."),
+        timeOfDay: import_zod36.z.string().optional().describe("Proposed 24-hour HH:MM run time, when the request named one."),
+        timezone: import_zod36.z.string().optional().describe("Proposed IANA timezone \u2014 from the request, or the account default when the request named a time without a zone."),
         rationale: import_zod36.z.string().optional().describe("Why these choices were made \u2014 read this before confirming."),
         error: import_zod36.z.string().optional()
       },
@@ -35394,7 +35468,7 @@ var init_memory_tool_schemas = __esm({
     ResumeScheduledActionSchema = {
       id: "resume-scheduled-action",
       upstreamName: "resumeScheduledActionTool",
-      description: "Resume a paused scheduled action. Its next run is computed fresh from now, not backfilled for time spent paused.",
+      description: "Resume a paused scheduled action for an active MCP Scraper Starter plan or higher. Its next run is computed fresh from now, not backfilled for time spent paused.",
       input: {
         id: import_zod36.z.string().min(1).describe("The scheduled action id.")
       },
@@ -35430,14 +35504,34 @@ var init_memory_tool_schemas = __esm({
         openWorldHint: false
       }
     };
+    SetScheduleDefaultsSchema = {
+      id: "set-schedule-defaults",
+      upstreamName: "setScheduleDefaultsTool",
+      description: "Set your default timezone for scheduled actions. Any schedule you later create with a time of day but no explicit timezone uses this default instead of UTC. Pass null to clear it back to UTC.",
+      input: {
+        defaultTimezone: import_zod36.z.string().nullable().describe('IANA timezone name, e.g. "America/Denver". null clears the default (new schedules fall back to UTC).')
+      },
+      output: {
+        ok: import_zod36.z.boolean(),
+        defaultTimezone: import_zod36.z.string().nullable().optional().describe("The default timezone now in effect; null means UTC."),
+        error: import_zod36.z.string().optional()
+      },
+      annotations: {
+        title: "Set Schedule Defaults",
+        readOnlyHint: false,
+        destructiveHint: false,
+        idempotentHint: true,
+        openWorldHint: false
+      }
+    };
     SetScheduleEntitlementSchema = {
       id: "set-schedule-entitlement",
       upstreamName: "setScheduleEntitlementTool",
-      description: "Admin-only scheduled-action credential provisioning and historical-row recovery. Rotate the encrypted delegated MCP Scraper API key without changing legacy enabled/quota fields by omitting those fields. Legacy entitlement values are not runtime access controls; paid-plan access is checked against MCP Scraper directly.",
+      description: "Admin-only scheduled-action credential provisioning and historical-row recovery. Use mcpScraperApiKey to rotate the encrypted delegated key without changing legacy enabled/quota fields. The former entitlement and quota values are retained only for migration compatibility and are not runtime access controls; paid-plan access is validated against MCP Scraper directly.",
       input: {
         granteeIdentity: import_zod36.z.string().min(1).describe("Identity whose scheduling entitlement is being set (e.g. an email)."),
         enabled: import_zod36.z.boolean().optional().describe("Historical entitlement value for migration/recovery only. Omit to preserve the stored value."),
-        quotaPerPeriod: import_zod36.z.number().optional().describe("Historical monthly execution quota retained only for migration compatibility. Omit to preserve the stored value."),
+        quotaPerPeriod: import_zod36.z.number().optional().describe("Historical monthly quota for migration/recovery only. Omit to preserve the stored value."),
         mcpScraperApiKey: import_zod36.z.string().optional().describe("The identity's mcp-scraper API key, stored encrypted, used to reach mcp-scraper tools during scheduled-action execution.")
       },
       output: {
@@ -35682,9 +35776,9 @@ var init_memory_tool_schemas = __esm({
     ListTagsSchema = {
       id: "list-memory-tags",
       upstreamName: "listTagsTool",
-      description: "List the live canonical tag vocabulary, aliases, usage counts, and per-vault distribution. Use this before choosing tags so existing concepts are reused instead of fragmented.",
+      description: "List the complete live canonical tag vocabulary, aliases, usage counts, and per-vault distribution. Always call this before proposing, resolving, or writing tags so the AI can reuse existing concepts and add only central, reusable concepts that are genuinely missing.",
       input: {
-        includeDeprecated: import_zod36.z.boolean().optional()
+        includeDeprecated: import_zod36.z.boolean().optional().describe("Include deprecated tags as well as active tags. Defaults true so the AI sees the complete vocabulary; pass false only for an active-only display.")
       },
       output: {
         ok: import_zod36.z.boolean(),
@@ -35696,7 +35790,7 @@ var init_memory_tool_schemas = __esm({
     ResolveTagsSchema = {
       id: "resolve-memory-tags",
       upstreamName: "resolveTagsTool",
-      description: "Resolve proposed concepts against the live tag vocabulary. Returns reuse, create, or omit. A new tag is allowed only when no equivalent exists and the caller marks the concept central and reusable.",
+      description: "Resolve proposed concepts against the live tag vocabulary. Always inspect the complete vocabulary with list-memory-tags first. Returns reuse, create, or omit; a new tag is appropriate only when no equivalent exists and the concept is central and reusable.",
       input: {
         candidates: import_zod36.z.array(import_zod36.z.object({ tag: import_zod36.z.string().min(1), central: import_zod36.z.boolean().optional(), reusable: import_zod36.z.boolean().optional(), description: import_zod36.z.string().optional() })).min(1).max(20)
       },
@@ -35710,7 +35804,7 @@ var init_memory_tool_schemas = __esm({
     UpsertTagSchema = {
       id: "upsert-memory-tag",
       upstreamName: "upsertTagTool",
-      description: "Define or curate one canonical tag, its meaning, aliases, and lifecycle. Use only after resolve-memory-tags returns create, or to merge/deprecate vocabulary intentionally. Requires write scope.",
+      description: "Define or curate one canonical tag, its meaning, aliases, and lifecycle. Inspect the complete vocabulary with list-memory-tags first; use only after resolve-memory-tags returns create, or to merge/deprecate vocabulary intentionally. Requires write scope.",
       input: {
         tag: import_zod36.z.string().min(1),
         description: import_zod36.z.string().optional(),
@@ -35866,17 +35960,17 @@ var init_memory_tool_schemas = __esm({
     ProvisionDefaultsSchema = {
       id: "provision-defaults",
       upstreamName: "provisionDefaultsTool",
-      description: "Provision the standard 13-vault memory structure (Ideas, Inspiration, Knowledge, Library, People, Communications, Calendar, Tasks, Projects, Issues, Improvement Log, Experiments, Sprint) for an identity. Idempotent \u2014 existing vaults are untouched. Optionally issues a fresh API key entitled to all 13. Requires admin scope.",
+      description: "Provision the standard 14-vault memory structure (Ideas, Inspiration, Knowledge, Library, People, Communications, Calendar, Tasks, Projects, Issues, Improvement Log, Experiments, Sprint, Skills) for an identity. Idempotent \u2014 existing vaults are untouched. Optionally issues a fresh API key entitled to all 14. Requires admin scope.",
       input: {
-        granteeIdentity: import_zod36.z.string().min(1).describe("Identity that should OWN the 13 default vaults (e.g. an email or user id)."),
-        issueKey: import_zod36.z.boolean().optional().describe("When true, also issue a new API key for the identity entitled to all 13 vaults and return its secret once. Default false."),
+        granteeIdentity: import_zod36.z.string().min(1).describe("Identity that should OWN the 14 default vaults (e.g. an email or user id)."),
+        issueKey: import_zod36.z.boolean().optional().describe("When true, also issue a new API key for the identity entitled to all 14 vaults and return its secret once. Default false."),
         plan: import_zod36.z.enum(["free", "pro", "team", "enterprise"]).optional().describe("Subscription plan carried by the issued key. Optional; defaults to free. Only used when issueKey is true.")
       },
       output: {
         ok: import_zod36.z.boolean().describe("True when provisioning succeeded; false on auth/scope error."),
         identity: import_zod36.z.string().optional().describe("Identity the vaults were provisioned for."),
-        created: import_zod36.z.array(import_zod36.z.string()).optional().describe("Vault names newly created on this call (empty when all 13 already existed)."),
-        vaults: import_zod36.z.array(import_zod36.z.string()).optional().describe("All 13 default vault names the identity now owns."),
+        created: import_zod36.z.array(import_zod36.z.string()).optional().describe("Vault names newly created on this call (empty when all 14 already existed)."),
+        vaults: import_zod36.z.array(import_zod36.z.string()).optional().describe("All 14 default vault names the identity now owns."),
         keyId: import_zod36.z.string().optional().describe("Stable id of the issued key. Present only when issueKey was true."),
         secret: import_zod36.z.string().optional().describe("The issued key secret \u2014 RETURNED ONCE. Present only when issueKey was true."),
         plan: import_zod36.z.string().optional().describe("Plan assigned to the issued key. Present only when issueKey was true."),
@@ -35913,9 +36007,9 @@ var init_memory_tool_schemas = __esm({
     VideoAnalyzeStartSchema = {
       id: "video-analyze-start",
       upstreamName: "videoAnalyzeStartTool",
-      description: "Start a deep async breakdown of a video: samples frames, transcribes audio, and runs parallel analyses (summary, pacing, WPM, topic outline, key points, hook analysis, visual style, replication recipe) into one saved report. Returns a runId immediately \u2014 poll video-analyze-status. Pass a direct video file URL (.mp4/.webm/.mov); platform URLs are not auto-resolved yet. Videos up to 1 hour.",
+      description: "Start a deep async breakdown of a video: samples frames, transcribes audio, and runs parallel analyses (summary, pacing, WPM, topic outline, key points, hook analysis, visual style, replication recipe) into one saved report. Returns a runId immediately \u2014 poll video-analyze-status. Accepts a YouTube, Facebook, Instagram, TikTok, or Vimeo URL directly (downloaded for you), or a direct video file URL (.mp4/.webm/.mov). Videos up to 1 hour.",
       input: {
-        sourceUrl: import_zod36.z.string().url().describe("Direct URL to the video file (.mp4/.webm/.mov/.gif)."),
+        sourceUrl: import_zod36.z.string().url().describe("A YouTube, Facebook, Instagram, TikTok, or Vimeo URL (downloaded automatically), or a direct video file URL (.mp4/.webm/.mov/.gif)."),
         intervalS: import_zod36.z.number().min(1).max(30).optional().describe("Preferred seconds between sampled frames (1-30). Default 2. For long videos the interval is automatically widened so the whole video is covered within the frame budget. Lower = denser sampling where the video is short enough to allow it."),
         maxFrames: import_zod36.z.number().int().min(1).max(480).optional().describe("Hard cap on frames analyzed (\u2264480). Default 120. Frames are spread across the whole duration; lowest sampling interval is 1 second, so short videos cannot use more frames than their length in seconds."),
         detail: import_zod36.z.enum(["fast", "standard", "deep"]).optional().describe("Analysis depth. Default standard."),
@@ -36052,6 +36146,7 @@ var init_memory_tool_schemas = __esm({
       GraphPathSchema,
       GraphUniverseSchema,
       LibraryIngestSchema,
+      BulkDeleteNotesSchema,
       DeleteNoteSchema,
       ExportSchema,
       GetSchema,
@@ -36070,6 +36165,7 @@ var init_memory_tool_schemas = __esm({
       ProposeScheduledActionSchema,
       ResumeScheduledActionSchema,
       RevokeScheduleLinkSchema,
+      SetScheduleDefaultsSchema,
       SetScheduleEntitlementSchema,
       CostUsageSchema,
       StorageUsageSchema,