@vellumai/assistant 0.8.6 → 0.8.7

package/src/daemon/message-types/messages.ts

@@ -1,12 +1,24 @@
 // User/assistant messages, tool results, confirmations, secrets, errors, and generation lifecycle.
 
+import type { AssistantActivityStateEvent } from "../../api/events/assistant-activity-state.js";
 import type { AssistantTextDeltaEvent } from "../../api/events/assistant-text-delta.js";
 import type { AssistantTurnStartEvent } from "../../api/events/assistant-turn-start.js";
+import type { ConfirmationRequestEvent } from "../../api/events/confirmation-request.js";
+import type { ErrorEvent } from "../../api/events/error.js";
+import type { InteractionResolvedEvent } from "../../api/events/interaction-resolved.js";
 import type { MessageCompleteEvent } from "../../api/events/message-complete.js";
+import type { MessageDequeuedEvent } from "../../api/events/message-dequeued.js";
+import type { MessageQueuedEvent } from "../../api/events/message-queued.js";
+import type { MessageQueuedDeletedEvent } from "../../api/events/message-queued-deleted.js";
+import type { MessageRequestCompleteEvent } from "../../api/events/message-request-complete.js";
+import type { QuestionRequestEvent } from "../../api/events/question-request.js";
+import type { SecretRequestEvent } from "../../api/events/secret-request.js";
+import type { ToolResultEvent } from "../../api/events/tool-result.js";
 import type { ToolUseStartEvent } from "../../api/events/tool-use-start.js";
+import type { TurnProfileAutoRoutedEvent } from "../../api/events/turn-profile-auto-routed.js";
+import type { UserMessageEchoEvent } from "../../api/events/user-message-echo.js";
 import type { ChannelId, InterfaceId } from "../../channels/types.js";
 import type { CommandIntent, UserMessageAttachment } from "./shared.js";
-import type { ToolActivityMetadata } from "./web-activity.js";
 
 // === Client → Server ===
 
@@ -29,7 +41,7 @@ export interface UserMessage {
   /** Structured command intent — bypasses text parsing when present. */
   commandIntent?: CommandIntent;
   /** Client-generated correlation nonce for echo dedup. See
-   *  `UserMessageEcho.clientMessageId` — the server echoes this value
+   *  `UserMessageEchoEvent.clientMessageId` — the server echoes this value
    *  back on the matching `user_message_echo` event. */
   clientMessageId?: string;
 }
@@ -58,24 +70,6 @@ export interface SuggestionRequest {
 
 // === Server → Client ===
 
-export interface UserMessageEcho {
-  type: "user_message_echo";
-  text: string;
-  conversationId?: string;
-  /** Database ID of the persisted user message, used by the originating
-   *  client to dedupe its optimistic row. Absent for synthetic echoes
-   *  (e.g. surface-action prompts) where no distinct user row is persisted. */
-  messageId?: string;
-  /** Server-generated request ID for the send. Allows correlation with
-   *  `message_queued` / `message_dequeued` events for the same turn. */
-  requestId?: string;
-  /** Client-generated correlation nonce from the HTTP POST body. Echoed
-   *  back so the originating client can dedupe its optimistic row even
-   *  if the SSE echo beats the 202 response. Absent for synthetic echoes
-   *  (e.g. surface-action prompts) that did not originate from a client POST. */
-  clientMessageId?: string;
-}
-
 export interface AssistantThinkingDelta {
   type: "assistant_thinking_delta";
   thinking: string;
@@ -122,233 +116,6 @@ export interface ToolInputDelta {
   messageId?: string;
 }
 
-export interface ToolResult {
-  type: "tool_result";
-  toolName: string;
-  result: string;
-  isError?: boolean;
-  diff?: {
-    filePath: string;
-    oldContent: string;
-    newContent: string;
-    isNewFile: boolean;
-  };
-  status?: string;
-  conversationId?: string;
-  /** Base64-encoded image data extracted from contentBlocks (e.g. browser_screenshot). @deprecated Use imageDataList. */
-  imageData?: string;
-  /** Base64-encoded image data extracted from contentBlocks (e.g. browser_screenshot, image generation). */
-  imageDataList?: string[];
-  /** The tool_use block ID for client-side correlation. */
-  toolUseId?: string;
-  /** Database ID of the assistant message that owns the parent tool_use
-   *  block. Same semantics as `AssistantTextDeltaEvent.messageId`. */
-  messageId?: string;
-  /** Risk level from the classifier ("low" | "medium" | "high" | "unknown"). */
-  riskLevel?: string;
-  /** Human-readable reason for the risk classification. */
-  riskReason?: string;
-  /** ID of the trust rule that matched this invocation (if any). */
-  matchedTrustRuleId?: string;
-  /** Whether the daemon is running in a containerized (Docker) environment. */
-  isContainerized?: boolean;
-  /**
-   * Display-only ladder of scope option labels for the rule editor
-   * (narrowest to broadest). The `pattern` here is regex-style and is
-   * NOT a valid trust rule pattern. Clients must use
-   * `riskAllowlistOptions` for the pattern that gets saved.
-   */
-  riskScopeOptions?: Array<{ pattern: string; label: string }>;
-  /**
-   * Allowlist options for the rule editor save path (narrowest to
-   * broadest). Each `pattern` is a Minimatch-glob compatible string —
-   * what the gateway actually matches against. Mirrors the
-   * `allowlistOptions` field on `ConfirmationRequest`. May be absent
-   * for tools whose classifier does not produce an allowlist (e.g.
-   * web-risk classifier, MCP tools without classifier coverage).
-   */
-  riskAllowlistOptions?: Array<{
-    label: string;
-    description: string;
-    pattern: string;
-  }>;
-  /** Directory scope ladder for the rule editor modal (narrowest to broadest). */
-  riskDirectoryScopeOptions?: Array<{ scope: string; label: string }>;
-  /** How the approval decision was reached: prompted, auto, blocked, or unknown (legacy). */
-  approvalMode?: string;
-  /** Why the approval decision was reached (stable enum for client display). */
-  approvalReason?: string;
-  /** Snapshot of the auto-approve threshold at execution time. */
-  riskThreshold?: string;
-  /** Structured activity metadata for rich client rendering. Optional; old
-   *  clients that key off `result` continue to work unchanged. */
-  activityMetadata?: ToolActivityMetadata;
-}
-
-export interface ConfirmationRequest {
-  type: "confirmation_request";
-  requestId: string;
-  toolName: string;
-  input: Record<string, unknown>;
-  riskLevel: string;
-  /** Human-readable reason for the risk classification (e.g. "Modifies remote repository state"). */
-  riskReason?: string;
-  /** Whether the daemon is running in a containerized (Docker) environment. */
-  isContainerized?: boolean;
-  executionTarget?: "sandbox" | "host";
-  allowlistOptions: Array<{
-    label: string;
-    description: string;
-    pattern: string;
-  }>;
-  scopeOptions: Array<{ label: string; scope: string }>;
-  directoryScopeOptions?: Array<{ scope: string; label: string }>;
-  diff?: {
-    filePath: string;
-    oldContent: string;
-    newContent: string;
-    isNewFile: boolean;
-  };
-  conversationId?: string;
-  /** When false, the client should hide "always allow" / trust-rule persistence affordances. */
-  persistentDecisionsAllowed?: boolean;
-  /** The tool_use block ID for client-side correlation with specific tool calls. */
-  toolUseId?: string;
-  /** ACP tool kind from the agent (e.g. "read", "edit", "execute"). Present only for ACP permission requests. */
-  acpToolKind?: string;
-  /** ACP permission options from the agent. Present only for ACP permission requests. Clients should use these to render the correct buttons. */
-  acpOptions?: Array<{
-    optionId: string;
-    name: string;
-    kind: "allow_once" | "allow_always" | "reject_once" | "reject_always";
-  }>;
-}
-
-export interface SecretRequest {
-  type: "secret_request";
-  requestId: string;
-  service: string;
-  field: string;
-  label: string;
-  description?: string;
-  placeholder?: string;
-  conversationId?: string;
-  /** Intended purpose of the credential (displayed to user). */
-  purpose?: string;
-  /** Tools allowed to use this credential. */
-  allowedTools?: string[];
-  /** Domains where this credential may be used. */
-  allowedDomains?: string[];
-  /** Whether one-time send override is available. */
-  allowOneTimeSend?: boolean;
-}
-
-export interface QuestionOption {
-  id: string;
-  label: string;
-  description?: string;
-}
-
-/**
- * One entry in a batched ask-question request.
- *
- * `id` is daemon-assigned (e.g. `q1`, `q2`...) — the LLM neither sees nor
- * supplies it. It exists so the client has a stable handle to post the
- * user's answer back against. See `QuestionRequest` for the batching
- * contract.
- */
-export interface QuestionEntry {
-  id: string;
-  question: string;
-  description?: string;
-  /** LLM-supplied options, capped at 4. The client always renders a fixed
-   *  5th "Type something else" slot wired to a free-text response — so this
-   *  array never represents the full choice set the user sees. */
-  options: QuestionOption[];
-  /** Optional placeholder shown in the free-text input. */
-  freeTextPlaceholder?: string;
-}
-
-/**
- * A single broadcast that carries either one or a small batch (≤5) of
- * clarifying questions. The whole batch is one card lifecycle on the client:
- * one render, one state machine, one response submission.
- *
- * Wire-compat plan: both shapes are populated on every broadcast.
- *  - `questions[]` is the canonical shape new clients should consume.
- *  - The flat `question` / `description` / `options` / `freeTextPlaceholder`
- *    fields mirror `questions[0]` for backwards compat with the existing
- *    web client, which keys off the flat fields. Once that client adopts
- *    `questions[]`, the flat fields can be dropped (separate cleanup).
- *
- * Daemon callers that don't supply a batch get a one-element `questions`
- * array synthesized from the flat fields.
- */
-export interface QuestionRequest {
-  type: "question_request";
-  requestId: string;
-  /** Batched-question payload. Always populated (single questions are sent
-   *  as a one-element array). Each entry's `id` is daemon-assigned. */
-  questions: QuestionEntry[];
-  /** Legacy: mirrors `questions[0].question`. Kept populated for clients
-   *  that haven't adopted the batched `questions[]` shape yet. */
-  question: string;
-  /** Legacy: mirrors `questions[0].description`. */
-  description?: string;
-  /** Legacy: mirrors `questions[0].options`. */
-  options: QuestionOption[];
-  /** Legacy: mirrors `questions[0].freeTextPlaceholder`. */
-  freeTextPlaceholder?: string;
-  conversationId?: string;
-  toolUseId?: string;
-}
-
-export interface ErrorMessage {
-  type: "error";
-  conversationId?: string;
-  requestId?: string;
-  code?: string;
-  message: string;
-  /** Categorizes the error so the client can offer contextual actions (e.g. "Send Anyway" for secret_blocked). */
-  category?: string;
-  /** Machine-readable conversation error category for clients that need source-aware recovery UI. */
-  errorCategory?: string;
-}
-
-export interface MessageQueued {
-  type: "message_queued";
-  conversationId: string;
-  requestId: string;
-  position: number;
-}
-
-export interface MessageDequeued {
-  type: "message_dequeued";
-  conversationId: string;
-  requestId: string;
-}
-
-/**
- * Request-level terminal signal for a user message lifecycle.
- *
- * Unlike `message_complete`, this does not imply the active assistant turn
- * has completed. It is used for paths that consume a request inline while a
- * separate in-flight turn may still be running.
- */
-export interface MessageRequestComplete {
-  type: "message_request_complete";
-  conversationId: string;
-  requestId: string;
-  /** True when an existing turn is still running after this request is finalized. */
-  runStillActive?: boolean;
-}
-
-export interface MessageQueuedDeleted {
-  type: "message_queued_deleted";
-  conversationId: string;
-  requestId: string;
-}
-
 export interface MessageSteered {
   type: "message_steered";
   conversationId: string;
@@ -382,89 +149,7 @@ export interface ConfirmationStateChanged {
   toolUseId?: string;
 }
 
-/**
- * Lifecycle states reported by `interaction_resolved`.
- *
- *  - `"approved"` / `"rejected"` — user-supplied verdict on a confirmation.
- *  - `"answered"` — user/client provided a response (secret value, question
- *    answer, host-proxy result).
- *  - `"cancelled"` — the interaction was torn down without a user response
- *    (timeout, abort, dispose, prompter shutdown).
- *  - `"superseded"` — invalidated by a newer event (auto-deny on enqueue, a
- *    fresh user message arriving while a confirmation was outstanding).
- */
-export type InteractionResolutionState =
-  | "approved"
-  | "rejected"
-  | "answered"
-  | "cancelled"
-  | "superseded";
-
-/**
- * Broadcast when a pending interaction (confirmation, secret, question,
- * host-proxy request) transitions to a resolved state. Clients use this to
- * drop attention/processing indicators without polling.
- */
-export interface InteractionResolved {
-  type: "interaction_resolved";
-  requestId: string;
-  /** Conversation id the interaction belongs to. */
-  conversationId: string;
-  state: InteractionResolutionState;
-  /** Kind of the resolved interaction (e.g. "confirmation", "secret", "host_bash"). */
-  kind: string;
-}
-
-/**
- * Server-side assistant activity lifecycle for thinking indicator placement.
- *
- * `activityVersion` is monotonically increasing per conversation. Clients must
- * ignore events with a version older than their current known version.
- */
-export interface AssistantActivityState {
-  type: "assistant_activity_state";
-  conversationId: string;
-  activityVersion: number;
-  phase:
-    | "idle"
-    | "thinking"
-    | "streaming"
-    | "tool_running"
-    | "awaiting_confirmation";
-  anchor: "assistant_turn" | "user_turn" | "global";
-  /** Active user request when available. */
-  requestId?: string;
-  reason:
-    | "message_dequeued"
-    | "thinking_delta"
-    | "first_text_delta"
-    | "tool_use_start"
-    | "preview_start"
-    | "tool_result_received"
-    | "confirmation_requested"
-    | "confirmation_resolved"
-    | "context_compacting"
-    | "message_complete"
-    | "generation_cancelled"
-    | "error_terminal";
-  /** Human-readable description of what the assistant is currently doing. */
-  statusText?: string;
-}
-
-/**
- * Emitted when the query complexity auto-router selects a non-default
- * profile for the current turn. Clients use this to show an inline
- * notification (e.g. "Using Quality for this response"). Only fires when
- * the router picks a profile — not when the user explicitly pinned one.
- */
-export interface TurnProfileAutoRouted {
-  type: "turn_profile_auto_routed";
-  conversationId: string;
-  /** Profile key (e.g. "quality-optimized"). */
-  profile: string;
-  /** Human-readable label (e.g. "Quality"). */
-  profileLabel: string;
-}
+export type { TurnProfileAutoRoutedEvent };
 
 /**
  * Broadcast to clients when a conversation's inference-profile override
@@ -520,7 +205,7 @@ export type _MessagesClientMessages =
   | SuggestionRequest;
 
 export type _MessagesServerMessages =
-  | UserMessageEcho
+  | UserMessageEchoEvent
   | AssistantTurnStartEvent
   | AssistantTextDeltaEvent
   | AssistantThinkingDelta
@@ -528,21 +213,21 @@ export type _MessagesServerMessages =
   | ToolUsePreviewStart
   | ToolOutputChunk
   | ToolInputDelta
-  | ToolResult
-  | ConfirmationRequest
-  | SecretRequest
-  | QuestionRequest
+  | ToolResultEvent
+  | ConfirmationRequestEvent
+  | SecretRequestEvent
+  | QuestionRequestEvent
   | MessageCompleteEvent
-  | ErrorMessage
-  | MessageQueued
-  | MessageDequeued
-  | MessageRequestComplete
-  | MessageQueuedDeleted
+  | ErrorEvent
+  | MessageQueuedEvent
+  | MessageDequeuedEvent
+  | MessageRequestCompleteEvent
+  | MessageQueuedDeletedEvent
   | MessageSteered
   | SuggestionResponse
   | TraceEvent
   | ConfirmationStateChanged
-  | AssistantActivityState
-  | TurnProfileAutoRouted
+  | AssistantActivityStateEvent
+  | TurnProfileAutoRoutedEvent
   | ConversationInferenceProfileUpdated
-  | InteractionResolved;
+  | InteractionResolvedEvent;

package/src/daemon/message-types/notifications.ts

@@ -1,34 +1,4 @@
-/** Broadcast to connected macOS clients when a notification should be displayed. */
-export interface NotificationIntent {
-  type: "notification_intent";
-  /** Delivery audit record ID so the client can correlate ack messages. */
-  deliveryId?: string;
-  sourceEventName: string;
-  title: string;
-  body: string;
-  /** Optional deep-link metadata so the client can navigate to the relevant context. */
-  deepLinkMetadata?: Record<string, unknown>;
-  /**
-   * When set, this notification is guardian-sensitive and should only be
-   * displayed by clients whose guardian identity matches this principal ID.
-   * Clients not bound to this guardian should ignore the notification.
-   */
-  targetGuardianPrincipalId?: string;
-  /**
-   * When true, the client must NOT post this intent to the OS notification
-   * surface (`UNUserNotificationCenter` on macOS). Non-banner side effects
-   * (guardian filtering, fallback dedup, mark-unseen + history catch-up on
-   * the paired conversation) still run. The home-feed inbox entry is
-   * written independently by `home-feed-side-effect.ts` and is unaffected
-   * by this flag.
-   *
-   * Set by the server based on `attentionHints.urgency`: true for
-   * `low`/`medium`, false for `high`/`critical`. The notification center
-   * is the always-on canonical inbox; the OS banner is reserved for
-   * signals the user opted into push for (urgency >= high).
-   */
-  silent?: boolean;
-}
+import type { NotificationIntentEvent } from "../../api/events/notification-intent.js";
 
 /** Server push — broadcast when a notification creates a new vellum conversation. */
 export interface NotificationConversationCreated {
@@ -105,5 +75,5 @@ export type _NotificationsClientMessages =
   | ConversationUnreadSignal;
 
 export type _NotificationsServerMessages =
-  | NotificationIntent
+  | NotificationIntentEvent
   | NotificationConversationCreated;

package/src/daemon/message-types/settings.ts

@@ -1,5 +1,7 @@
 // Client settings: daemon-pushed configuration updates to connected clients.
 
+import type { AvatarUpdatedEvent } from "../../api/events/avatar-updated.js";
+
 // === Client → Server ===
 
 /** Request from a conversation or client to change the voice activation key. */
@@ -27,13 +29,6 @@ export interface ClientSettingsUpdate {
   value: string;
 }
 
-/** Sent by the daemon after the avatar image has been regenerated and saved to disk. */
-export interface AvatarUpdated {
-  type: "avatar_updated";
-  /** Absolute path to the updated avatar image file. */
-  avatarPath: string;
-}
-
 /** Sent by the daemon when workspace config.json changes on disk. */
 export interface ConfigChanged {
   type: "config_changed";
@@ -60,7 +55,7 @@ export type _SettingsClientMessages =
   | GenerateAvatarRequest;
 export type _SettingsServerMessages =
   | ClientSettingsUpdate
-  | AvatarUpdated
+  | AvatarUpdatedEvent
   | ConfigChanged
   | SoundsConfigUpdated
   | GenerateAvatarResponse;

package/src/daemon/message-types/skills.ts

@@ -89,6 +89,7 @@ interface SlimSkillBase {
   emoji?: string;
   kind: "bundled" | "installed" | "catalog";
   status: "enabled" | "disabled" | "available";
+  category: string;
 }
 
 interface VellumSlimSkill extends SlimSkillBase {
@@ -159,6 +160,7 @@ interface SkillDetailBase {
   emoji?: string;
   kind: "bundled" | "installed" | "catalog";
   status: "enabled" | "disabled" | "available";
+  category: string;
 }
 
 interface VellumSkillDetail extends SkillDetailBase {

package/src/daemon/message-types/surfaces.ts

@@ -188,6 +188,8 @@ interface UiSurfaceShowBase {
   messageId?: string;
   /** When `true`, clicking an action does not dismiss the surface — the client keeps the card visible and only marks the clicked `actionId` as spent so siblings remain clickable. */
   persistent?: boolean;
+  /** Id of the tool call that produced this surface (the `ui_show` proxy tool). Lets the client gate app previews on the tool result's arrival rather than whole-turn streaming state. */
+  toolCallId?: string;
 }
 
 export interface UiSurfaceShowCard extends UiSurfaceShowBase {

package/src/daemon/message-types/sync.ts

@@ -1,11 +1,20 @@
 import { z } from "zod";
 
+import {
+  type SyncChangedEvent,
+  SyncChangedEventSchema,
+} from "../../api/events/sync-changed.js";
+
+export { type SyncChangedEvent, SyncChangedEventSchema };
+
 export const SYNC_TAGS = {
   assistantAvatar: "assistant:self:avatar",
   assistantIdentity: "assistant:self:identity",
+  assistantIdentityIntro: "assistant:self:identity-intro",
   assistantConfig: "assistant:self:config",
   assistantSounds: "assistant:self:sounds",
   assistantSchedules: "assistant:self:schedules",
+  appsList: "apps:list",
   conversationsList: "conversations:list",
   featureFlagsClient: "feature-flags:client",
   featureFlagsAssistant: "feature-flags:assistant",
@@ -23,29 +32,8 @@ export type SyncInvalidationTag =
   | ConversationSyncInvalidationTag
   | (string & {});
 
-export interface SyncChangedMessage {
-  type: "sync_changed";
-  tags: SyncInvalidationTag[];
-  /**
-   * Optional identifier of the client that originated the change. When set,
-   * the server fan-out and clients themselves can suppress self-echoes so
-   * the originating tab/process doesn't reinvalidate its own cache off its
-   * own mutation. Daemon-internal emits (agent loop, FS watcher, cron) leave
-   * this unset so the event fans out to every subscriber as before.
-   */
-  originClientId?: string;
-}
-
 export const SyncInvalidationTagSchema = z.string().min(1);
 
-export const SyncChangedMessageSchema = z
-  .object({
-    type: z.literal("sync_changed"),
-    tags: z.array(SyncInvalidationTagSchema).min(1),
-    originClientId: z.string().min(1).optional(),
-  })
-  .strict();
-
 export function conversationMessagesSyncTag(
   conversationId: string,
 ): ConversationSyncInvalidationTag {
@@ -61,15 +49,14 @@ export function conversationMetadataSyncTag(
 export function buildSyncChangedMessage(
   tags: SyncInvalidationTag[],
   originClientId?: string,
-): SyncChangedMessage {
+): SyncChangedEvent {
   const dedupedTags = Array.from(new Set(tags));
   const trimmedOrigin = originClientId?.trim();
-  const parsed = SyncChangedMessageSchema.parse({
+  return SyncChangedEventSchema.parse({
     type: "sync_changed",
     tags: dedupedTags,
     ...(trimmedOrigin ? { originClientId: trimmedOrigin } : {}),
   });
-  return parsed as SyncChangedMessage;
 }
 
-export type _SyncInvalidationServerMessages = SyncChangedMessage;
+export type _SyncInvalidationServerMessages = SyncChangedEvent;

package/src/daemon/message-types/workspace.ts

@@ -1,5 +1,7 @@
 // Workspace file, identity, and tool permission types.
 
+import type { IdentityChangedEvent } from "../../api/events/identity-changed.js";
+
 // === Client → Server ===
 
 export interface WorkspaceFilesListRequest {
@@ -117,16 +119,6 @@ export interface ToolNamesListResponse {
   schemas?: Record<string, ToolInputSchema>;
 }
 
-/** Server push — broadcast when IDENTITY.md changes on disk. */
-export interface IdentityChanged {
-  type: "identity_changed";
-  name: string;
-  role: string;
-  personality: string;
-  emoji: string;
-  home: string;
-}
-
 // --- Domain-level union aliases (consumed by the barrel file) ---
 
 export type _WorkspaceClientMessages =
@@ -142,4 +134,4 @@ export type _WorkspaceServerMessages =
   | IdentityGetResponse
   | ToolPermissionSimulateResponse
   | ToolNamesListResponse
-  | IdentityChanged;
+  | IdentityChangedEvent;