@vellumai/assistant 0.8.0 → 0.8.2

package/src/config/bundled-skills/contacts/SKILL.md

@@ -15,27 +15,7 @@ Manage the user's contacts, relationship graph, access control (trusted contacts
 
 ## Contact Management
 
-### Create or update a contact
-
-Create a new contact or update an existing one in the relationship graph. Use this to track people the user interacts with across channels.
-
-```bash
-assistant contacts upsert --display-name "<name>" --notes "<notes>" --channels '<json_array>' --json
-```
-
-To update an existing contact, include the `--id` flag.
-
-Required flags:
-
-- `--display-name` -- the contact's name
-
-Optional flags:
-
-- `--id` -- contact ID to update (omit to create new, or auto-match by channel address)
-- `--notes` -- free-text notes about this contact (e.g. relationship, communication preferences, response expectations)
-- `--role` -- contact role: `contact` or `guardian` (default: `contact`)
-- `--contact-type` -- contact type: `human` or `assistant` (default: `human`)
-- `--channels` -- JSON array of channel objects, each with `type`, `address`, and optional `isPrimary`, `externalUserId`, `externalChatId`, `status`, `policy`
+> **Contact writes are guardian-only.** You can list, search, and merge contacts, but creating contacts and granting access happen through the guardian-facing flows: the **Contacts tab in the assistant web UI** for new contacts and channel verification, and `assistant contacts invites create` for invite-based onboarding. The LLM does **not** have a direct create/update tool — when a user asks for one, walk them to the guardian flow instead.
 
 ### Search contacts
 
@@ -116,25 +96,12 @@ The response contains `{ ok: true, contacts: [...] }` where each contact has:
 
 ### Allow a user (add trusted contact)
 
-Use this when the user wants to grant someone access to message the assistant. **Always confirm with the user before executing this action.**
-
-Ask the user: _"I'll add [name/identifier] on [channel] as an allowed contact. Should I proceed?"_
-
-```bash
-assistant contacts upsert --display-name "<display_name>" --channels '[{"type":"<channel>","address":"<user_id>","externalUserId":"<user_id>","status":"active","policy":"allow"}]' --json
-```
-
-Required flags:
+You can't add a trusted contact directly — contact writes are guardian-only. When the user wants to grant someone access, walk them through one of these guardian paths:
 
-- `--display-name` -- human-readable name for the contact
-- `--channels` -- at least one channel entry with:
-  - `type` -- the channel type (e.g., `telegram`)
-  - `address` -- the channel-specific identifier
-  - `externalUserId` -- the user's ID on that channel (or `externalChatId` for chat-based channels)
-  - `status` -- set to `"active"` for immediate access
-  - `policy` -- set to `"allow"` to grant messaging access
+1. **Web UI** — open the Contacts tab in the assistant dashboard, click "Add contact", and verify the channel via the outbound verification flow (SMS code, Telegram message, etc.).
+2. **Invite link** — for channels that support it (Telegram, email, etc.), create an invite with `assistant contacts invites create --source-channel <channel> --contact-id <existing_contact_id>` (the contact must already exist).
 
-If the user provides a name but not an external ID, explain that you need the channel-specific user ID or chat ID to create the contact entry. For Telegram, this is a numeric user ID.
+If the user insists on adding someone, tell them: _"I can't add trusted contacts directly — that's a guardian-only action. You can add them from the Contacts tab in the assistant dashboard, or I can create an invite link for an existing contact."_
 
 ### Revoke a user (remove access)
 
@@ -169,7 +136,7 @@ Replace `<channel_id>` with the channel's `id` from the contact's `channels` arr
 
 Invite links let the guardian share a link or code that automatically grants access when used. Telegram invites use a deep link; voice invites use a phone number + numeric code; email, WhatsApp, and Slack invites use a 6-digit code that the invitee sends to the assistant on the respective channel.
 
-**Every invite must be bound to a contact.** Before creating an invite, look up the contact with `assistant contacts list` or create one with `assistant contacts upsert`, then pass the contact's `id` via the required `--contact-id` flag.
+**Every invite must be bound to a contact.** Before creating an invite, look up the contact with `assistant contacts list` and pass the contact's `id` via the required `--contact-id` flag. **You cannot create new contacts** — if the target contact doesn't exist yet, tell the user to add them from the **Contacts tab in the assistant web UI** first, then come back to create the invite.
 
 ### Create a Telegram invite link
 
@@ -216,7 +183,7 @@ echo "$INVITE_URL"
 Required flags:
 
 - `--source-channel` -- must be `telegram`
-- `--contact-id` -- the ID of the contact this invite is for. Look up or create the contact first with `assistant contacts list` or `assistant contacts upsert`.
+- `--contact-id` -- the ID of the contact this invite is for. Look up the contact first with `assistant contacts list`. New contacts must be added via the Contacts tab in the assistant web UI; you cannot create them here.
 
 Optional flags:
 
@@ -256,7 +223,7 @@ assistant contacts invites create --source-channel phone --contact-id "<contact_
 Required flags:
 
 - `--source-channel` -- must be `phone`
-- `--contact-id` -- the ID of the contact this invite is for. Look up or create the contact first with `assistant contacts list` or `assistant contacts upsert`.
+- `--contact-id` -- the ID of the contact this invite is for. Look up the contact first with `assistant contacts list`. New contacts must be added via the Contacts tab in the assistant web UI; you cannot create them here.
 - `--expected-external-user-id` -- the invitee's phone number in E.164 format (e.g., `+15551234567`)
 - `--friend-name` -- the invitee's display name (e.g., "Mom", "Dr. Smith"). Used during the voice verification call to personalize the experience.
 - `--guardian-name` -- the guardian's display name (e.g., "Alex"). Used during the voice verification call so the invitee knows who invited them.
@@ -302,7 +269,7 @@ assistant contacts invites create --source-channel email --contact-id "<contact_
 Required flags:
 
 - `--source-channel` -- must be `email`
-- `--contact-id` -- the ID of the contact this invite is for. Look up or create the contact first with `assistant contacts list` or `assistant contacts upsert`.
+- `--contact-id` -- the ID of the contact this invite is for. Look up the contact first with `assistant contacts list`. New contacts must be added via the Contacts tab in the assistant web UI; you cannot create them here.
 
 The response contains `{ ok: true, invite: { id, token, inviteCode, guardianInstruction, channelHandle, ... } }`.
 
@@ -333,7 +300,7 @@ assistant contacts invites create --source-channel whatsapp --contact-id "<conta
 Required flags:
 
 - `--source-channel` -- must be `whatsapp`
-- `--contact-id` -- the ID of the contact this invite is for. Look up or create the contact first with `assistant contacts list` or `assistant contacts upsert`.
+- `--contact-id` -- the ID of the contact this invite is for. Look up the contact first with `assistant contacts list`. New contacts must be added via the Contacts tab in the assistant web UI; you cannot create them here.
 
 The response contains `{ ok: true, invite: { id, token, inviteCode, guardianInstruction, channelHandle?, ... } }`.
 
@@ -366,7 +333,7 @@ assistant contacts invites create --source-channel slack --contact-id "<contact_
 Required flags:
 
 - `--source-channel` -- must be `slack`
-- `--contact-id` -- the ID of the contact this invite is for. Look up or create the contact first with `assistant contacts list` or `assistant contacts upsert`.
+- `--contact-id` -- the ID of the contact this invite is for. Look up the contact first with `assistant contacts list`. New contacts must be added via the Contacts tab in the assistant web UI; you cannot create them here.
 
 The response follows the same shape as email and WhatsApp invites (`inviteCode`, `guardianInstruction`, `channelHandle`).
 
@@ -497,7 +464,7 @@ Each channel has:
 
 **"Who can message me?"** -- List all contacts with `assistant contacts list --json`, present active channels as a formatted list.
 
-**"Add my friend to Telegram"** -- Ask for their Telegram user ID (numeric) and display name, confirm, then create a contact with `assistant contacts upsert` including a channel entry with `policy: "allow"` and `status: "active"`.
+**"Add my friend to Telegram"** -- You can't add trusted contacts directly (guardian-only action). Tell the user to add the contact from the Contacts tab in the assistant dashboard, where they can enter the Telegram user ID and verify the channel. Once the contact exists, you can create an invite with `assistant contacts invites create --source-channel telegram --contact-id <contact_id>`.
 
 **"Remove [name]'s access"** -- List contacts with `assistant contacts list --json` to find them, identify the channel to revoke, confirm the revocation, then run `assistant contacts channels update-status <channel_id> --status revoked --json`.
 

package/src/config/bundled-skills/contacts/TOOLS.json

@@ -1,63 +1,6 @@
 {
   "version": 1,
   "tools": [
-    {
-      "name": "contact_upsert",
-      "description": "Create or update a contact in the relationship graph. Use this to track people the user interacts with across channels (email, Slack, etc.).",
-      "category": "contacts",
-      "risk": "low",
-      "input_schema": {
-        "type": "object",
-        "properties": {
-          "id": {
-            "type": "string",
-            "description": "Contact ID to update. Omit to create a new contact (or auto-match by channel address)."
-          },
-          "display_name": {
-            "type": "string",
-            "description": "Display name for the contact"
-          },
-          "notes": {
-            "type": "string",
-            "description": "Free-text notes about this contact (e.g. relationship, communication preferences, response expectations)"
-          },
-          "channels": {
-            "type": "array",
-            "description": "Communication channels for this contact",
-            "items": {
-              "type": "object",
-              "properties": {
-                "type": {
-                  "type": "string",
-                  "enum": [
-                    "email",
-                    "slack",
-                    "whatsapp",
-                    "phone",
-                    "telegram",
-                    "discord",
-                    "other"
-                  ],
-                  "description": "Channel type"
-                },
-                "address": {
-                  "type": "string",
-                  "description": "Channel address (email address, Slack handle, phone number, etc.)"
-                },
-                "is_primary": {
-                  "type": "boolean",
-                  "description": "Whether this is the primary channel for this type"
-                }
-              },
-              "required": ["type", "address"]
-            }
-          }
-        },
-        "required": ["display_name"]
-      },
-      "executor": "tools/contact-upsert.ts",
-      "execution_target": "host"
-    },
     {
       "name": "contact_search",
       "description": "Search for contacts by name, channel address, or other criteria. Returns matching contacts with their channel information.",

package/src/config/bundled-skills/document/SKILL.md

@@ -18,13 +18,33 @@ Create and edit long-form documents using the built-in rich text editor. Documen
 
 - **document_create** - Opens a new document editor with an optional title and initial Markdown content. Returns a `surface_id` for subsequent updates.
 - **document_update** - Updates content in an open document editor by `surface_id`. Supports `replace` (overwrite) and `append` (add to end) modes.
+- **document_read** - Reads the current content of a document by `surface_id` when it belongs to the current conversation, or when the current actor is the guardian/local user. Use to verify content before editing.
+- **document_list** - Lists documents. Without `query`, lists the current conversation's documents. With `query`, searches by title; guardian/local users can search across conversations, while other actors are scoped to the current conversation.
+- **document_delete** - Deletes a document by `surface_id`. Use to clean up unwanted documents.
 
-## Workflow
+## Retrieving existing documents
+
+When the user asks to see, open, or pull up a document:
+
+1. Check the `<active_documents>` block in your context — it lists all documents in this conversation with their `surface_id` and title. If the document is there, call `document_read` with its `surface_id`. Done in one call.
+2. If the document is NOT in `<active_documents>`, call `document_list` with a `query` matching the document title. For guardian/local users, this searches across previous conversations and sessions.
+3. Once you have the `surface_id`, call `document_read` to retrieve the content.
+
+**Never** search the filesystem, conversation history, or archives to find a document. Always use `document_list` with a `query`.
+
+## Creating a new document
 
 1. **Create the document**: Call `document_create` with a title (inferred from the request). Call the tool immediately, not after conversational preamble.
-2. **Write content in Markdown**: Use proper structure (`#` for titles, `##` for sections), **bold**, *italic*, code blocks, tables, lists, blockquotes as appropriate.
+2. **Write content in Markdown**: Use proper structure (`#` for titles, `##` for sections), **bold**, _italic_, code blocks, tables, lists, blockquotes as appropriate.
 3. **CRITICAL - Stream content in chunks**: Call `document_update` MULTIPLE times, not just once. Break content into logical chunks (paragraphs, sections, or every 200-300 words). Call `document_update` with `mode: "append"` for EACH chunk separately. The user experiences real-time content appearing as you write.
-4. **Respond to edits**: When the user requests changes via the docked chat, use `document_update` with `replace` for full rewrites or `append` for additions.
+
+## Editing an existing document
+
+When the user requests changes to a document:
+
+1. Find the `surface_id` from the `<active_documents>` context block.
+2. Use `document_update` with the existing `surface_id` — do NOT call `document_create` again.
+3. Use `mode: "replace"` for full rewrites or `mode: "append"` for additions.
 
 ## Usage Notes
 

package/src/config/bundled-skills/document/TOOLS.json

@@ -48,6 +48,59 @@
       },
       "executor": "tools/document-update.ts",
       "execution_target": "host"
+    },
+    {
+      "name": "document_read",
+      "description": "Read the current content of a document by its surface_id when it belongs to the current conversation, or when the current actor is the guardian/local user. Use this to verify document state before making edits.",
+      "category": "document",
+      "risk": "low",
+      "input_schema": {
+        "type": "object",
+        "properties": {
+          "surface_id": {
+            "type": "string",
+            "description": "The ID of the document to read"
+          }
+        },
+        "required": ["surface_id"]
+      },
+      "executor": "tools/document-read.ts",
+      "execution_target": "host"
+    },
+    {
+      "name": "document_list",
+      "description": "List documents. Without a query, lists documents in the current conversation. With a query, searches documents by title; guardian/local users can search across conversations, while other actors are scoped to the current conversation.",
+      "category": "document",
+      "risk": "low",
+      "input_schema": {
+        "type": "object",
+        "properties": {
+          "query": {
+            "type": "string",
+            "description": "Search documents by title. Omit to list only the current conversation's documents."
+          }
+        }
+      },
+      "executor": "tools/document-list.ts",
+      "execution_target": "host"
+    },
+    {
+      "name": "document_delete",
+      "description": "Delete a document by its surface_id.",
+      "category": "document",
+      "risk": "low",
+      "input_schema": {
+        "type": "object",
+        "properties": {
+          "surface_id": {
+            "type": "string",
+            "description": "The ID of the document to delete"
+          }
+        },
+        "required": ["surface_id"]
+      },
+      "executor": "tools/document-delete.ts",
+      "execution_target": "host"
     }
   ]
 }

package/src/config/bundled-skills/document/tools/document-delete.ts

@@ -0,0 +1,12 @@
+import { executeDocumentDelete } from "../../../../tools/document/document-tool.js";
+import type {
+  ToolContext,
+  ToolExecutionResult,
+} from "../../../../tools/types.js";
+
+export async function run(
+  input: Record<string, unknown>,
+  context: ToolContext,
+): Promise<ToolExecutionResult> {
+  return executeDocumentDelete(input, context);
+}

package/src/config/bundled-skills/document/tools/document-list.ts

@@ -0,0 +1,12 @@
+import { executeDocumentList } from "../../../../tools/document/document-tool.js";
+import type {
+  ToolContext,
+  ToolExecutionResult,
+} from "../../../../tools/types.js";
+
+export async function run(
+  input: Record<string, unknown>,
+  context: ToolContext,
+): Promise<ToolExecutionResult> {
+  return executeDocumentList(input, context);
+}

package/src/config/bundled-skills/document/tools/document-read.ts

@@ -0,0 +1,12 @@
+import { executeDocumentRead } from "../../../../tools/document/document-tool.js";
+import type {
+  ToolContext,
+  ToolExecutionResult,
+} from "../../../../tools/types.js";
+
+export async function run(
+  input: Record<string, unknown>,
+  context: ToolContext,
+): Promise<ToolExecutionResult> {
+  return executeDocumentRead(input, context);
+}

package/src/config/bundled-skills/messaging/tools/messaging-archive-by-sender.ts

@@ -1,13 +1,9 @@
-import { emitFeedEvent } from "../../../../home/emit-feed-event.js";
 import type {
   ToolContext,
   ToolExecutionResult,
 } from "../../../../tools/types.js";
-import { getLogger } from "../../../../util/logger.js";
 import { err, getProviderConnection, ok, resolveProvider } from "./shared.js";
 
-const log = getLogger("messaging-archive-by-sender");
-
 export async function run(
   input: Record<string, unknown>,
   context: ToolContext,
@@ -50,14 +46,6 @@ export async function run(
     }
 
     const summary = `Archived ${result.archived} message(s) matching query: ${query}`;
-    void emitFeedEvent({
-      source: "gmail",
-      title: "Messages Archived",
-      summary,
-      dedupKey: `email-archive:${Date.now()}`,
-    }).catch((err) => {
-      log.warn({ err }, "Failed to emit email archive feed event");
-    });
     if (result.truncated) {
       return ok(
         `${summary}\n\nNote: this operation was capped at 5000 messages. Additional messages matching the query may remain in the inbox. Run the command again to archive more.`,

package/src/config/bundled-skills/messaging/tools/messaging-send.ts

@@ -1,7 +1,6 @@
 import { readFile } from "node:fs/promises";
 import { basename } from "node:path";
 
-import { emitFeedEvent } from "../../../../home/emit-feed-event.js";
 import {
   addMessage,
   getConversation,
@@ -150,14 +149,6 @@ export async function run(
             ccList.length > 0
               ? `To: ${toList.join(", ")}; Cc: ${ccList.join(", ")}`
               : `To: ${toList.join(", ")}`;
-          void emitFeedEvent({
-            source: "gmail",
-            title: "Email Draft Created",
-            summary: `Drafted reply to ${recipientSummary}.`,
-            dedupKey: `email-draft:${draft.id}`,
-          }).catch((err) => {
-            log.warn({ err }, "Failed to emit email draft feed event");
-          });
           return ok(
             `Gmail draft created with ${attachments.length} attachment(s): ${filenames} (Draft ID: ${draft.id}). ${recipientSummary}. Review in Gmail Drafts, then tell me to send it or send it yourself.`,
           );
@@ -178,14 +169,6 @@ export async function run(
           ccList.length > 0
             ? `To: ${toList.join(", ")}; Cc: ${ccList.join(", ")}`
             : `To: ${toList.join(", ")}`;
-        void emitFeedEvent({
-          source: "gmail",
-          title: "Email Draft Created",
-          summary: `Drafted reply to ${recipientSummary}.`,
-          dedupKey: `email-draft:${draft.id}`,
-        }).catch((err) => {
-          log.warn({ err }, "Failed to emit email draft feed event");
-        });
         return ok(
           `Gmail draft created (ID: ${draft.id}). ${recipientSummary}. Review in Gmail Drafts, then tell me to send it or send it yourself.`,
         );
@@ -212,14 +195,6 @@ export async function run(
         const draft = await createDraftRaw(gmailConn, raw, threadId);
 
         const filenames = attachments.map((a) => a.filename).join(", ");
-        void emitFeedEvent({
-          source: "gmail",
-          title: "Email Draft Created",
-          summary: "Created an email draft.",
-          dedupKey: `email-draft:${draft.id}`,
-        }).catch((err) => {
-          log.warn({ err }, "Failed to emit email draft feed event");
-        });
         return ok(
           `Gmail draft created with ${attachments.length} attachment(s): ${filenames} (Draft ID: ${draft.id}). Review in Gmail Drafts, then tell me to send it or send it yourself.`,
         );
@@ -236,14 +211,6 @@ export async function run(
         undefined,
         threadId,
       );
-      void emitFeedEvent({
-        source: "gmail",
-        title: "Email Draft Created",
-        summary: "Created an email draft.",
-        dedupKey: `email-draft:${draft.id}`,
-      }).catch((err) => {
-        log.warn({ err }, "Failed to emit email draft feed event");
-      });
       return ok(
         `Gmail draft created (ID: ${draft.id}). Review it in your Gmail Drafts, then tell me to send it or send it yourself from Gmail.`,
       );
@@ -257,31 +224,6 @@ export async function run(
       assistantId: context.assistantId,
     });
 
-    const sendSummary =
-      provider.id === "slack"
-        ? "Sent a Slack message."
-        : provider.id === "telegram"
-          ? "Sent a Telegram message."
-          : "Sent an email.";
-    void emitFeedEvent({
-      source:
-        provider.id === "slack"
-          ? "slack"
-          : provider.id === "telegram"
-            ? "telegram"
-            : "gmail",
-      title:
-        provider.id === "slack"
-          ? "Slack Message Sent"
-          : provider.id === "telegram"
-            ? "Telegram Message Sent"
-            : "Email Sent",
-      summary: sendSummary,
-      dedupKey: `message-sent:${result.id}`,
-    }).catch((err) => {
-      log.warn({ err }, "Failed to emit message send feed event");
-    });
-
     const threadSuffix = result.threadId
       ? `, "thread_id": "${result.threadId}"`
       : "";

package/src/config/bundled-skills/skill-management/SKILL.md

@@ -18,6 +18,6 @@ Manage the lifecycle of custom managed skills in `{workspaceDir}/skills`.
 ## Capabilities
 
 - **Scaffold** a new managed skill with YAML frontmatter and markdown body
-- **Delete** an existing managed skill and remove it from the SKILLS.md index
+- **Delete** an existing managed skill directory
 
-Skills created via `scaffold_managed_skill` become available for `skill_load` immediately.
+Skills created via `scaffold_managed_skill` become available for `skill_load` when a valid top-level `SKILL.md` is written under the skill directory.

package/src/config/bundled-skills/skill-management/TOOLS.json

@@ -3,7 +3,7 @@
   "tools": [
     {
       "name": "scaffold_managed_skill",
-      "description": "Create or update a managed skill in {workspaceDir}/skills. The skill becomes available for skill_load immediately. Never persist a skill without explicit user consent. Before persisting, test the snippet: write to a temp file with bash and run with `bun run /tmp/vellum-eval/snippet.ts`. Iterate up to 3 attempts, then ask the user. Clean up temp files after. Do not use file_write for temp files outside the working directory. After a skill is written, the next turn may run in a recreated conversation due to file-watcher eviction - continue normally.",
+      "description": "Create or update a managed skill in {workspaceDir}/skills. The skill becomes available for skill_load when a valid top-level SKILL.md is written under the skill directory. Never persist a skill without explicit user consent. Before persisting, test the snippet: write to a temp file with bash and run with `bun run /tmp/vellum-eval/snippet.ts`. Iterate up to 3 attempts, then ask the user. Clean up temp files after. Do not use file_write for temp files outside the working directory. After a skill is written, the next turn may run in a recreated conversation due to file-watcher eviction - continue normally.",
       "category": "skills",
       "risk": "high",
       "input_schema": {
@@ -33,14 +33,14 @@
             "type": "boolean",
             "description": "Whether to overwrite an existing skill with the same ID (default: false)."
           },
-          "add_to_index": {
-            "type": "boolean",
-            "description": "Whether to add the skill to SKILLS.md index (default: true)."
-          },
           "includes": {
             "type": "array",
             "items": { "type": "string" },
             "description": "Optional list of child skill IDs that this skill includes (metadata only, no auto-activation)."
+          },
+          "add_to_index": {
+            "type": "boolean",
+            "description": "Deprecated no-op compatibility field. Skills are discovered from top-level SKILL.md files."
           }
         },
         "required": ["skill_id", "name", "description", "body_markdown"]
@@ -50,7 +50,7 @@
     },
     {
       "name": "delete_managed_skill",
-      "description": "Delete a managed skill from {workspaceDir}/skills and remove it from the SKILLS.md index. Never delete a skill without explicit user confirmation. After deletion, the next turn may run in a recreated conversation due to file-watcher eviction - continue normally.",
+      "description": "Delete a managed skill directory from {workspaceDir}/skills. Never delete a skill without explicit user confirmation. After deletion, the next turn may run in a recreated conversation due to file-watcher eviction - continue normally.",
       "category": "skills",
       "risk": "high",
       "input_schema": {
@@ -62,7 +62,7 @@
           },
           "remove_from_index": {
             "type": "boolean",
-            "description": "Whether to remove the skill from SKILLS.md index (default: true)."
+            "description": "Deprecated no-op compatibility field. Skill deletion does not edit SKILLS.md."
           }
         },
         "required": ["skill_id"]

package/src/config/bundled-tool-registry.ts

@@ -49,10 +49,12 @@ import * as computerUseWait from "./bundled-skills/computer-use/tools/computer-u
 // ── contacts ───────────────────────────────────────────────────────────────────
 import * as contactMerge from "./bundled-skills/contacts/tools/contact-merge.js";
 import * as contactSearch from "./bundled-skills/contacts/tools/contact-search.js";
-import * as contactUpsert from "./bundled-skills/contacts/tools/contact-upsert.js";
 import * as googleContacts from "./bundled-skills/contacts/tools/google-contacts.js";
 // ── document ───────────────────────────────────────────────────────────────────
 import * as documentCreate from "./bundled-skills/document/tools/document-create.js";
+import * as documentDelete from "./bundled-skills/document/tools/document-delete.js";
+import * as documentList from "./bundled-skills/document/tools/document-list.js";
+import * as documentRead from "./bundled-skills/document/tools/document-read.js";
 import * as documentUpdate from "./bundled-skills/document/tools/document-update.js";
 // ── followups ──────────────────────────────────────────────────────────────────
 import * as followupCreate from "./bundled-skills/followups/tools/followup-create.js";
@@ -163,13 +165,15 @@ export const bundledToolRegistry = new Map<string, SkillToolScript>([
   ["computer-use:tools/computer-use-respond.ts", computerUseRespond],
 
   // contacts
-  ["contacts:tools/contact-upsert.ts", contactUpsert],
   ["contacts:tools/contact-search.ts", contactSearch],
   ["contacts:tools/contact-merge.ts", contactMerge],
   ["contacts:tools/google-contacts.ts", googleContacts],
 
   // document
   ["document:tools/document-create.ts", documentCreate],
+  ["document:tools/document-delete.ts", documentDelete],
+  ["document:tools/document-list.ts", documentList],
+  ["document:tools/document-read.ts", documentRead],
   ["document:tools/document-update.ts", documentUpdate],
 
   // followups

package/src/config/feature-flag-registry.json

@@ -9,6 +9,14 @@
       "description": "Automatically trigger conversation analysis on the same cadence as memory extraction (batch threshold, idle debounce, end-of-conversation). The analysis agent has full tool access and writes back to memory and skills without user approval.",
       "defaultEnabled": false
     },
+    {
+      "id": "memory-retrospective",
+      "scope": "assistant",
+      "key": "memory-retrospective",
+      "label": "Memory retrospective pass",
+      "description": "Run a focused, memory-only retrospective during active conversations and at conversation lifecycle. The retrospective agent re-reads the messages added since the last successful run, dedupes against memory/archive/, and calls `remember` for what wasn't captured in the moment. Enabling this also relaxes the in-conversation pressure to call `remember` (lighter PKB reminder + relaxed tool description) so the assistant can stay present and trust the retrospective as the backstop.",
+      "defaultEnabled": false
+    },
     {
       "id": "user-hosted-enabled",
       "scope": "client",
@@ -193,6 +201,14 @@
       "description": "Route embedding requests through the platform runtime proxy using Vellum-managed Gemini credentials when available",
       "defaultEnabled": false
     },
+    {
+      "id": "bookmarks",
+      "scope": "client",
+      "key": "bookmarks",
+      "label": "Message Bookmarks",
+      "description": "Show the bookmark icon on messages and enable the Bookmarks tab in Settings",
+      "defaultEnabled": false
+    },
     {
       "id": "fork-from-message",
       "scope": "client",
@@ -265,6 +281,14 @@
       "description": "Enable the app-control skill (per-app screenshot + raw input bypassing AX tree)",
       "defaultEnabled": false
     },
+    {
+      "id": "species-migration",
+      "scope": "assistant",
+      "key": "species-migration",
+      "label": "Species Migration",
+      "description": "Enable the Species Migration skill for migrating from OpenClaw, Hermes, Manus, and other assistant species into Vellum.",
+      "defaultEnabled": false
+    },
     {
       "id": "analyze-conversation",
       "scope": "assistant",
@@ -278,7 +302,39 @@
       "scope": "assistant",
       "key": "pro-plan-adjust",
       "label": "Pro Plan Adjust",
-      "description": "Show the rich Plan card (current plan, features, Manage/Upgrade CTA) at the top of the macOS Settings → Billing tab.",
+      "description": "Show the rich Plan card (current plan, features, Manage/Upgrade CTA) at the top of the macOS Settings \u2192 Billing tab.",
+      "defaultEnabled": false
+    },
+    {
+      "id": "external-plugins",
+      "scope": "assistant",
+      "key": "external-plugins",
+      "label": "External Plugins",
+      "description": "Enable the external-plugin install path: the `assistant plugins` CLI subcommand and the declarative external-plugin loader convention (`<workspaceDir>/plugins/<name>/` directories containing `package.json` plus interface dirs). Gates an unstable surface that may change shape before stabilizing.",
+      "defaultEnabled": false
+    },
+    {
+      "id": "provider-zai",
+      "scope": "assistant",
+      "key": "provider-zai",
+      "label": "z.ai Provider",
+      "description": "Enable the z.ai (Zhipu AI) provider and its GLM models in the provider picker and model selection UI",
+      "defaultEnabled": false
+    },
+    {
+      "id": "provider-deepseek",
+      "scope": "assistant",
+      "key": "provider-deepseek",
+      "label": "DeepSeek Provider",
+      "description": "Enable the DeepSeek direct API provider and its models (V4 Pro, V4 Flash) in the provider picker and model selection UI",
+      "defaultEnabled": false
+    },
+    {
+      "id": "provider-minimax",
+      "scope": "assistant",
+      "key": "provider-minimax",
+      "label": "MiniMax Provider",
+      "description": "Enable the MiniMax direct API provider and its models (M2.7, M2.5, M2.1, M2, and highspeed variants) in the provider picker and model selection UI",
       "defaultEnabled": false
     }
   ]