@vellumai/assistant 0.10.8 → 0.10.9-dev.202607152225.ec6fdbf

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

@@ -125,7 +125,7 @@
     },
     {
       "name": "messaging_send",
-      "description": "Send a message, reply to a thread, or create a draft. On Gmail, always creates a draft for review. Supports replies (via thread_id), attachments (via attachment_paths, Gmail only), and all messaging platforms.",
+      "description": "Send a message, reply to a thread, or create a draft. On Gmail, always creates a draft for review. Supports replies (via thread_id), attachments (via attachment_paths, Gmail and Outlook), and all messaging platforms.",
       "category": "messaging",
       "risk": "high",
       "input_schema": {
@@ -160,7 +160,7 @@
             "items": {
               "type": "string"
             },
-            "description": "Absolute file paths of attachments to include (Gmail only)"
+            "description": "Absolute file paths of attachments to include (Gmail and Outlook)"
           },
           "thread_id": {
             "type": "string",

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

@@ -1,6 +1,7 @@
 import { readFile } from "node:fs/promises";
 import { basename } from "node:path";
 
+import type { OutboundAttachment } from "../../../../messaging/provider-types.js";
 import {
   createDraft,
   createDraftRaw,
@@ -19,7 +20,7 @@ import type {
   ToolExecutionResult,
 } from "../../../../tools/types.js";
 import { getLogger } from "../../../../util/logger.js";
-import { guessMimeType } from "./gmail-mime-helpers.js";
+import { guessMimeType } from "../../../../util/mime-type.js";
 import {
   err,
   extractEmail,
@@ -32,6 +33,20 @@ import {
 
 const log = getLogger("messaging-send");
 
+/** Read attachment files from disk into in-memory parts for outbound sending. */
+async function readAttachments(paths: string[]): Promise<OutboundAttachment[]> {
+  return Promise.all(
+    paths.map(async (filePath) => ({
+      filename: basename(filePath),
+      mimeType: guessMimeType(filePath),
+      data: await readFile(filePath),
+    })),
+  );
+}
+
+/** Email providers that accept file attachments on outbound sends. */
+const ATTACHMENT_CAPABLE_PLATFORMS = new Set(["gmail", "outlook"]);
+
 export async function run(
   input: Record<string, unknown>,
   context: ToolContext,
@@ -54,9 +69,12 @@ export async function run(
   try {
     const provider = await resolveProvider(platform);
 
-    // Non-Gmail platforms: reject attachment_paths
-    if (provider.id !== "gmail" && attachmentPaths?.length) {
-      return err("Attachments are only supported on Gmail.");
+    // Reject attachments on platforms that can't carry them (e.g. Telegram, WhatsApp).
+    if (
+      attachmentPaths?.length &&
+      !ATTACHMENT_CAPABLE_PLATFORMS.has(provider.id)
+    ) {
+      return err("Attachments are only supported on Gmail and Outlook.");
     }
 
     const account = input.account as string | undefined;
@@ -125,14 +143,7 @@ export async function run(
 
         // With attachments: build multipart MIME for threaded reply
         if (attachmentPaths?.length) {
-          const attachments = await Promise.all(
-            attachmentPaths.map(async (filePath) => {
-              const data = await readFile(filePath);
-              const filename = basename(filePath);
-              const mimeType = guessMimeType(filePath);
-              return { filename, mimeType, data };
-            }),
-          );
+          const attachments = await readAttachments(attachmentPaths);
 
           const raw = buildMultipartMime({
             to: toList.join(", "),
@@ -176,14 +187,7 @@ export async function run(
 
       // With attachments: build multipart MIME and use createDraftRaw
       if (attachmentPaths?.length) {
-        const attachments = await Promise.all(
-          attachmentPaths.map(async (filePath) => {
-            const data = await readFile(filePath);
-            const filename = basename(filePath);
-            const mimeType = guessMimeType(filePath);
-            return { filename, mimeType, data };
-          }),
-        );
+        const attachments = await readAttachments(attachmentPaths);
 
         const raw = buildMultipartMime({
           to: conversationId,
@@ -217,10 +221,14 @@ export async function run(
     }
 
     // Non-Gmail platforms
+    const attachments = attachmentPaths?.length
+      ? await readAttachments(attachmentPaths)
+      : undefined;
     const result = await provider.sendMessage(conn, conversationId, text, {
       subject,
       inReplyTo,
       threadId,
+      attachments,
       assistantId: context.assistantId,
     });
 

package/src/config/bundled-skills/phone-calls/references/CONFIG.md

@@ -10,7 +10,7 @@ All call-related settings can be managed via `assistant config`:
 | `calls.userConsultTimeoutSeconds`           | How long to wait for user answers                                                                                                                                                                                                         | `120` (2 min)                                                                                            |
 | `calls.disclosure.enabled`                  | Whether the AI announces itself at call start                                                                                                                                                                                             | `true`                                                                                                   |
 | `calls.disclosure.text`                     | The disclosure message spoken at call start                                                                                                                                                                                               | `"At the very beginning of the call, introduce yourself as an assistant calling on behalf of my human."` |
-| `llm.callSites.callAgent.model`             | Override LLM model for call orchestration                                                                                                                                                                                                 | _(unset — falls back to `llm.default.model`)_                                                            |
+| `llm.callSites.callAgent.model`             | Override LLM model for call orchestration                                                                                                                                                                                                 | _(unset — falls back to the resolved call-site default)_                                                 |
 | `calls.callerIdentity.allowPerCallOverride` | Allow per-call caller identity selection                                                                                                                                                                                                  | `true`                                                                                                   |
 | `calls.callerIdentity.userNumber`           | E.164 phone number for user-number mode                                                                                                                                                                                                   | _(empty)_                                                                                                |
 | `calls.voice.language`                      | Language code for TTS and transcription                                                                                                                                                                                                   | `en-US`                                                                                                  |
@@ -22,12 +22,12 @@ All call-related settings can be managed via `assistant config`:
 
 During phone calls the assistant synthesizes audio server-side via the configured provider's API, transcodes it to mu-law, and streams it to Twilio over the media stream. Each provider's catalog entry declares its playback format:
 
-| Provider     | Playback format | Description                                                                                                                       |
-| ------------ | --------------- | ---------------------------------------------------------------------------------------------------------------------------------- |
-| `elevenlabs` | `pcm`           | Synthesizes PCM audio via the ElevenLabs API.                                                                                     |
-| `fish-audio` | `wav`           | Synthesizes WAV audio via Fish Audio's HTTP API. Requires a configured `referenceId`.                                             |
-| `deepgram`   | `pcm`           | Synthesizes PCM audio via Deepgram's HTTP API. Uses the same API key as Deepgram speech-to-text.                                  |
-| `xai`        | `pcm`           | Streams PCM audio via xAI's WebSocket endpoint.                                                                                   |
+| Provider     | Playback format | Description                                                                                      |
+| ------------ | --------------- | ------------------------------------------------------------------------------------------------ |
+| `elevenlabs` | `pcm`           | Synthesizes PCM audio via the ElevenLabs API.                                                    |
+| `fish-audio` | `wav`           | Synthesizes WAV audio via Fish Audio's HTTP API. Requires a configured `referenceId`.            |
+| `deepgram`   | `pcm`           | Synthesizes PCM audio via Deepgram's HTTP API. Uses the same API key as Deepgram speech-to-text. |
+| `xai`        | `pcm`           | Streams PCM audio via xAI's WebSocket endpoint.                                                  |
 
 Calls require a media-stream-playable TTS provider with a working API key. When the configured provider is not playable (unsupported format or missing credentials), the assistant falls back to a credentialed playable provider rather than letting the call go silent; if none exists, calls are blocked up front with a setup-required message.
 

package/src/config/bundled-skills/phone-calls/references/TROUBLESHOOTING.md

@@ -16,18 +16,18 @@ First check whether this is a managed/platform assistant:
 assistant platform status --json
 ```
 
-If it reports an available platform assistant, do not install or start ngrok for Twilio. The gateway should use Velay, and `velayTunnel.connected` should become `true` after registration. If this is a local/self-hosted assistant without Velay, run the **public-ingress** skill to set up ngrok or another custom tunnel and configure `ingress.publicBaseUrl`.
+If it reports an available platform assistant, do not install or start ngrok for Twilio. The gateway should use Velay; run `assistant gateway status --json` and confirm it reports a `tunnel` URL (a `{ "tunnel": "..." }` object, not `{}`) after registration. If this is a local/self-hosted assistant without Velay, run the **public-ingress** skill to set up ngrok or another custom tunnel and configure `ingress.publicBaseUrl`.
 
 ## Call fails immediately after initiating
 
 - Check that the phone number is in E.164 format
 - Verify Twilio credentials are correct (wrong auth token causes API errors)
 - On trial accounts, ensure the destination number is verified
-- Check that the configured tunnel is still running. For ngrok, use `curl -s http://127.0.0.1:4040/api/tunnels`. For Velay, run `assistant platform status --json` and confirm `velayTunnel.connected` is `true`, or check gateway logs for `Velay tunnel registered`.
+- Check that the configured tunnel is still running. For ngrok, use `curl -s http://127.0.0.1:4040/api/tunnels`. For Velay, run `assistant gateway status --json` and confirm it reports a `tunnel` URL (not `{}`), or check gateway logs for `Velay tunnel registered`.
 
 ## Call connects but no audio / one-way audio
 
-- The media-stream WebSocket may not be connecting. If you are using ngrok or a custom tunnel, check that `ingress.publicBaseUrl` is correct and the tunnel is active. If you are using Velay, check `assistant platform status --json`; a 503 from `https://velay.../<assistant-id>/...` usually means the assistant tunnel is not connected or the gateway did not complete the WebSocket open, not that ngrok is required.
+- The media-stream WebSocket may not be connecting. If you are using ngrok or a custom tunnel, check that `ingress.publicBaseUrl` is correct and the tunnel is active. If you are using Velay, check `assistant gateway status --json`; a 503 from `https://velay.../<assistant-id>/...` usually means the assistant tunnel is not connected or the gateway did not complete the WebSocket open, not that ngrok is required.
 - Verify the assistant is running
 
 ## "Number not eligible for caller identity"

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

@@ -175,7 +175,6 @@ Use `syntax` + `expression` to specify the schedule type explicitly, or just `ex
 
 - **When the user specifies a name for the schedule, use it exactly as given.** Do not paraphrase, embellish, or generate a descriptive name.
 - Use `schedule_create` for both recurring automation ("every day at 9am") and one-time reminders ("remind me at 3pm").
-- For task tracking ("add to my tasks", "add to my queue"), use task_list_add instead.
 - `fire_at` must be a strict ISO 8601 timestamp with timezone offset or Z (e.g. `2025-03-15T09:00:00-05:00`).
 
 ### Anchored & Ambiguous Relative Time

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

@@ -3,7 +3,7 @@
   "tools": [
     {
       "name": "schedule_create",
-      "description": "Create a scheduled automation. Always provide a concise description of what the schedule does and why. For recurring: provide expression (cron/RRULE). For one-time: provide fire_at (ISO 8601 timestamp). ONLY use this when the user explicitly wants something to run on a schedule (e.g. \"every day at 9am\", \"remind me tomorrow at 3pm\", \"weekly on Mondays\"). Do NOT use this for \"add to my tasks\" or \"add to my queue\" - use task_list_add for those requests instead.",
+      "description": "Create a scheduled automation. Always provide a concise description of what the schedule does and why. For recurring: provide expression (cron/RRULE). For one-time: provide fire_at (ISO 8601 timestamp). ONLY use this when the user explicitly wants something to run on a schedule (e.g. \"every day at 9am\", \"remind me tomorrow at 3pm\", \"weekly on Mondays\").",
       "category": "schedule",
       "risk": "medium",
       "input_schema": {
@@ -120,7 +120,7 @@
     },
     {
       "name": "schedule_list",
-      "description": "List scheduled automations (recurring cron/RRULE schedules and one-time scheduled events), or show details and recent runs for a specific one. For the user's task queue, use task_list_show instead.",
+      "description": "List scheduled automations (recurring cron/RRULE schedules and one-time scheduled events), or show details and recent runs for a specific one.",
       "category": "schedule",
       "risk": "low",
       "input_schema": {

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

@@ -3,7 +3,7 @@
   "tools": [
     {
       "name": "voice_config_update",
-      "description": "Update a voice configuration setting. Use tts_provider to switch the global TTS provider (valid providers: elevenlabs, fish-audio, deepgram, xai — defined by the provider catalog). Use tts_voice_id to set the ElevenLabs voice ID, fish_audio_reference_id for Fish Audio voice reference. Deepgram uses the same API key as speech-to-text and requires no additional voice config. Changes persist to services.tts config and take effect immediately.",
+      "description": "Update a voice configuration setting. Use tts_provider to switch the global TTS provider (valid providers: elevenlabs, fish-audio, deepgram, xai — defined by the provider catalog). Use tts_voice_id to set the ElevenLabs voice ID, fish_audio_reference_id for Fish Audio voice reference. Deepgram uses the same API key as speech-to-text and requires no additional voice config. Use stt_mode / tts_mode to switch between \"your-own\" (bring your own provider API key) and \"managed\" (Vellum-managed speech, billed to your organization; requires a Vellum platform connection via 'assistant platform connect'). Changes persist to services.stt / services.tts config and take effect immediately.",
       "category": "system",
       "risk": "low",
       "input_schema": {
@@ -15,13 +15,15 @@
               "activation_key",
               "conversation_timeout",
               "fish_audio_reference_id",
+              "stt_mode",
+              "tts_mode",
               "tts_provider",
               "tts_voice_id"
             ],
-            "description": "The voice setting to change. tts_provider selects the global TTS provider from the provider catalog (valid providers: elevenlabs, fish-audio, deepgram, xai). tts_voice_id sets the ElevenLabs voice. fish_audio_reference_id sets the Fish Audio voice reference. Deepgram uses its default model and shares the STT API key."
+            "description": "The voice setting to change. tts_provider selects the global TTS provider from the provider catalog (valid providers: elevenlabs, fish-audio, deepgram, xai). tts_voice_id sets the ElevenLabs voice. fish_audio_reference_id sets the Fish Audio voice reference. Deepgram uses its default model and shares the STT API key. stt_mode / tts_mode select \"your-own\" or \"managed\" speech for each service."
           },
           "value": {
-            "description": "The new value for the setting. For tts_provider: one of elevenlabs, fish-audio, deepgram, xai. For tts_voice_id: an alphanumeric ElevenLabs voice ID. For fish_audio_reference_id: a Fish Audio voice reference ID. For conversation_timeout: seconds (5, 10, 15, 30, or 60). For activation_key: key identifier string."
+            "description": "The new value for the setting. For tts_provider: one of elevenlabs, fish-audio, deepgram, xai. For tts_voice_id: an alphanumeric ElevenLabs voice ID. For fish_audio_reference_id: a Fish Audio voice reference ID. For conversation_timeout: seconds (5, 10, 15, 30, or 60). For activation_key: key identifier string. For stt_mode / tts_mode: \"your-own\" or \"managed\"."
           }
         }
       },

package/src/config/bundled-skills/settings/tools/voice-config-update.ts

@@ -1,10 +1,12 @@
 import { normalizeActivationKey } from "../../../../daemon/handlers/config-voice.js";
+import { managedSpeechAvailable } from "../../../../platform/managed-speech.js";
 import type {
   ToolContext,
   ToolExecutionResult,
 } from "../../../../tools/types.js";
 import { listCatalogProviderIds } from "../../../../tts/provider-catalog.js";
 import {
+  getConfig,
   invalidateConfigCache,
   loadRawConfig,
   saveRawConfig,
@@ -15,24 +17,30 @@ import { VALID_CONVERSATION_TIMEOUTS } from "../../../schemas/elevenlabs.js";
 /**
  * Valid voice config settings and their UserDefaults key mappings.
  *
- * All config paths are canonical (`services.tts.*`).
+ * All config paths are canonical (`services.tts.*` / `services.stt.*`).
+ * Settings without a userDefaultsKey are daemon-config-only and are not
+ * broadcast to the desktop client.
  */
+type VoiceSettingMeta = { userDefaultsKey?: string; type: "string" | "number" };
+
 const VOICE_SETTINGS = {
   activation_key: {
     userDefaultsKey: "pttActivationKey",
-    type: "string" as const,
+    type: "string",
   },
   conversation_timeout: {
     userDefaultsKey: "voiceConversationTimeoutSeconds",
-    type: "number" as const,
+    type: "number",
   },
-  tts_provider: { userDefaultsKey: "ttsProvider", type: "string" as const },
-  tts_voice_id: { userDefaultsKey: "ttsVoiceId", type: "string" as const },
+  tts_provider: { userDefaultsKey: "ttsProvider", type: "string" },
+  tts_voice_id: { userDefaultsKey: "ttsVoiceId", type: "string" },
   fish_audio_reference_id: {
     userDefaultsKey: "fishAudioReferenceId",
-    type: "string" as const,
+    type: "string",
   },
-} as const;
+  stt_mode: { type: "string" },
+  tts_mode: { type: "string" },
+} satisfies Record<string, VoiceSettingMeta>;
 
 type VoiceSettingName = keyof typeof VOICE_SETTINGS;
 
@@ -46,8 +54,12 @@ const FRIENDLY_NAMES: Record<VoiceSettingName, string> = {
   tts_provider: "TTS provider",
   tts_voice_id: "ElevenLabs voice",
   fish_audio_reference_id: "Fish Audio voice",
+  stt_mode: "Speech-to-text mode",
+  tts_mode: "Text-to-speech mode",
 };
 
+const VALID_SPEECH_MODES = ["your-own", "managed"] as const;
+
 function validateSetting(
   setting: string,
   value: unknown,
@@ -117,6 +129,21 @@ function validateSetting(
       }
       return { ok: true, coerced: value.trim() };
     }
+    case "stt_mode":
+    case "tts_mode": {
+      if (
+        typeof value !== "string" ||
+        !VALID_SPEECH_MODES.includes(
+          value.trim() as (typeof VALID_SPEECH_MODES)[number],
+        )
+      ) {
+        return {
+          ok: false,
+          error: `${setting} must be one of: ${VALID_SPEECH_MODES.join(", ")}`,
+        };
+      }
+      return { ok: true, coerced: value.trim() };
+    }
     case "fish_audio_reference_id": {
       if (typeof value !== "string" || value.trim().length === 0) {
         return {
@@ -159,13 +186,25 @@ export async function run(
     return { content: `Error: ${validation.error}`, isError: true };
   }
 
-  const meta = VOICE_SETTINGS[setting as VoiceSettingName];
+  if (
+    (setting === "stt_mode" || setting === "tts_mode") &&
+    validation.coerced === "managed" &&
+    !(await managedSpeechAvailable())
+  ) {
+    return {
+      content:
+        "Error: managed speech requires a Vellum platform connection. Run 'assistant platform connect' first.",
+      isError: true,
+    };
+  }
+
+  const meta: VoiceSettingMeta = VOICE_SETTINGS[setting as VoiceSettingName];
   const friendlyName = FRIENDLY_NAMES[setting as VoiceSettingName];
 
   // Send client_settings_update message to write to UserDefaults.
   // Always stringify the value — Swift's ClientSettingsUpdate.value is typed
   // as String, so a bare JSON number would fail to decode.
-  if (context.sendToClient) {
+  if (context.sendToClient && meta.userDefaultsKey) {
     context.sendToClient({
       type: "client_settings_update",
       key: meta.userDefaultsKey,
@@ -212,10 +251,45 @@ export async function run(
     invalidateConfigCache();
   }
 
+  if (setting === "stt_mode") {
+    setNestedValue(raw, "services.stt.mode", validation.coerced);
+    // SttServiceSchema requires `provider` whenever the stt object exists and
+    // forbids provider "vellum" outside managed mode, so writing `mode` alone
+    // (or keeping "vellum" when switching to your-own) would make the saved
+    // config invalid.
+    const currentProvider = getConfig().services.stt.provider;
+    setNestedValue(
+      raw,
+      "services.stt.provider",
+      currentProvider === "vellum" && validation.coerced === "your-own"
+        ? "deepgram"
+        : currentProvider,
+    );
+    saveRawConfig(raw);
+    invalidateConfigCache();
+  }
+
+  if (setting === "tts_mode") {
+    setNestedValue(raw, "services.tts.mode", validation.coerced);
+    // TtsServiceSchema forbids provider "vellum" outside managed mode, so
+    // switching to your-own must also replace a vellum provider.
+    if (
+      validation.coerced === "your-own" &&
+      getConfig().services.tts.provider === "vellum"
+    ) {
+      setNestedValue(raw, "services.tts.provider", "elevenlabs");
+    }
+    saveRawConfig(raw);
+    invalidateConfigCache();
+  }
+
+  const broadcastNote = meta.userDefaultsKey
+    ? " The change has been broadcast to the desktop client."
+    : "";
   return {
     content: `${friendlyName} updated to ${JSON.stringify(
       validation.coerced,
-    )}. The change has been broadcast to the desktop client.`,
+    )}.${broadcastNote}`,
     isError: false,
   };
 }

package/src/config/bundled-skills/transcribe/tools/transcribe-media.test.ts

@@ -7,13 +7,6 @@ import type { ToolContext } from "../../../../tools/types.js";
 // Mocks — must be declared before the subject import
 // ---------------------------------------------------------------------------
 
-mock.module("../../../../util/logger.js", () => ({
-  getLogger: () =>
-    new Proxy({} as Record<string, unknown>, {
-      get: () => () => {},
-    }),
-}));
-
 let mockTranscriber: BatchTranscriber | null = null;
 
 mock.module("../../../../providers/speech-to-text/resolve.js", () => ({

package/src/config/call-site-defaults.ts

@@ -2,9 +2,9 @@ import { type LLMCallSite } from "./schemas/llm.js";
 
 type CallSiteDefaultConfig = {
   /**
-   * Named profile the call site resolves to. Omit to inherit the workspace
-   * default config (`llm.default`, with the active profile applied) — used for
-   * call sites that must resolve to a credentialed provider on every install
+   * Named profile the call site resolves to. Omit to fall through to the
+   * balanced intent resolved through `llm.defaultProvider` — used for call
+   * sites that must resolve to a credentialed provider on every install
    * rather than pinning a profile that may be unavailable.
    */
   profile?: string;

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

@@ -9,18 +9,6 @@
       "description": "Enable user-hosted onboarding flow",
       "defaultEnabled": false
     },
-    {
-      "id": "pre-chat-onboarding-experiment-2026-06-06",
-      "scope": "client",
-      "key": "pre-chat-onboarding-experiment-2026-06-06",
-      "label": "Pre-chat Onboarding Experiment 2026-06-06",
-      "description": "Pre-chat onboarding experiment with control and variant-a arms. Control shows the full funnel; variant-a shows the condensed pared-down flow.",
-      "defaultEnabled": "control",
-      "values": [
-        "control",
-        "variant-a"
-      ]
-    },
     {
       "id": "experiment-activation-flow-2026-06-03",
       "scope": "both",
@@ -28,11 +16,7 @@
       "label": "Activation Flow Experiment 2026-06-03",
       "description": "Multivariate activation-flow experiment. control = standard flow; variant-a = activation rail; personal-page = new sign-up-page variant (front-end only). Targeted via LaunchDarkly.",
       "defaultEnabled": "control",
-      "values": [
-        "control",
-        "variant-a",
-        "personal-page"
-      ]
+      "values": ["control", "variant-a", "personal-page"]
     },
     {
       "id": "local-docker-enabled",
@@ -50,14 +34,6 @@
       "description": "Enable the A2A (Agent-to-Agent) channel for inter-assistant communication via the open A2A protocol",
       "defaultEnabled": false
     },
-    {
-      "id": "trace-collection",
-      "scope": "assistant",
-      "key": "trace-collection",
-      "label": "Trace Collection",
-      "description": "Gate per-turn conversation trace (user/assistant/tool-call/tool-response) collection. The daemon attaches a trace to its turn telemetry only when this flag AND the owner's share_diagnostics consent are both on. Defaults off (fail-closed) \u2014 the live value is delivered from LaunchDarkly.",
-      "defaultEnabled": false
-    },
     {
       "id": "settings-developer-nav",
       "scope": "assistant",
@@ -122,6 +98,14 @@
       "description": "Show the live voice conversation button in the composer",
       "defaultEnabled": false
     },
+    {
+      "id": "voice-triage-escalate",
+      "scope": "assistant",
+      "key": "voice-triage-escalate",
+      "label": "Voice Triage & Escalate",
+      "description": "Front phone-call turns with a fast model that answers simple turns and escalates tricky ones to the quality model, cutting perceived latency",
+      "defaultEnabled": false
+    },
     {
       "id": "fast-mode",
       "scope": "assistant",
@@ -178,14 +162,6 @@
       "description": "Show a live HUD in the top-right of the conversation with scroll geometry, velocity, update rate, and anchor-preserver activity. Developer diagnostic for investigating scroll jank.",
       "defaultEnabled": false
     },
-    {
-      "id": "coding-agents-panel",
-      "scope": "client",
-      "key": "coding-agents-panel",
-      "label": "Coding Agents Panel",
-      "description": "Gate native macOS Coding Agents panel entry points and inline ACP session deep links.",
-      "defaultEnabled": false
-    },
     {
       "id": "compaction-playground",
       "scope": "assistant",
@@ -311,7 +287,7 @@
       "scope": "assistant",
       "key": "memory-concept-graph",
       "label": "Memory concept graph",
-      "description": "Gates the Obsidian-style memory concept graph on the assistant identity page and the backend-agnostic /memory-graph + /memory-graph-node routes that feed it. When off, the identity page shows the skills constellation instead. Only produces a graph on memory-v3 backends. Default off.",
+      "description": "Gates the Obsidian-style memory concept graph on its own Memory tab (/assistant/memory) and the backend-agnostic /memory-graph + /memory-graph-node routes that feed it. When off, the Memory tab is hidden; the identity page keeps the skills constellation. Only produces a graph on memory-v3 backends. Default off.",
       "defaultEnabled": false
     },
     {
@@ -322,22 +298,6 @@
       "description": "On the first conversation, send a natural self-introduction (e.g. \"Hi Vela, I'm alex. Nice to meet you.\") on the user's behalf and route it through real LLM inference, instead of serving the canned first greeting. Names come from the onboarding context; falls back to the canned greeting when no name is known. Exposed to clients so pre-hatch onboarding can compute the source-of-truth initial message, and to the assistant so older clients can still be gated server-side. See assistant/src/daemon/first-greeting.ts (buildSelfIntroMessage).",
       "defaultEnabled": false
     },
-    {
-      "id": "research-onboarding",
-      "scope": "client",
-      "key": "research-onboarding",
-      "label": "Research onboarding (spike)",
-      "description": "Spike: replace pre-chat onboarding with a web-research \"here's what I know about you\" flow, including a Google Calendar \"Let's chat tomorrow\" step shown over the streaming research. Gates the /assistant/onboarding/research route and the research mock harness. Off by default; enable locally via the feature-flags panel.",
-      "defaultEnabled": false
-    },
-    {
-      "id": "personality-onboarding",
-      "scope": "client",
-      "key": "personality-onboarding",
-      "label": "Personality onboarding step (spike)",
-      "description": "Spike: show the \"Create my personality\" trait-slider step in the research-onboarding flow (between the pitch and the integration step). On continue it applies the sliders to the assistant's persona on a throwaway side conversation. When off, the step is skipped entirely. Off by default; enable locally via the feature-flags panel.",
-      "defaultEnabled": false
-    },
     {
       "id": "channel-trust-floors",
       "scope": "assistant",
@@ -403,20 +363,12 @@
       "defaultEnabled": false
     },
     {
-      "id": "skill-creation-card",
-      "scope": "both",
-      "key": "skill-creation-card",
-      "label": "Skill creation card",
-      "description": "Insert an in-chat card into the source conversation when the memory retrospective authors a new skill, and render it in the web client.",
-      "defaultEnabled": false
-    },
-    {
-      "id": "override-or-default-resolution",
+      "id": "chat-credential-reveal",
       "scope": "assistant",
-      "key": "override-or-default-resolution",
-      "label": "Override-or-Default Profile Resolution",
-      "description": "Resolve inference profiles as single complete overrides (or the call site's default intent through the default provider) instead of the legacy deep-merge cascade. Ships enabled; disable as a kill switch to restore merge semantics.",
-      "defaultEnabled": true
+      "key": "chat-credential-reveal",
+      "label": "Chat Credential Reveal",
+      "description": "Persist chat secret redactions as sentinel markers instead of the legacy HTML marker, enriched with the vault identity when the value byte-matches a credential revealed this turn. Also gates the `credentials reveal --for-chat` channel and the live-stream swap that replaces an echoed reveal plaintext with its sentinel before emission. Redaction itself is NOT gated: a route-proven reveal plaintext stays out of persisted rows in both modes. Client chip rendering is gated on daemon version, not on this flag.",
+      "defaultEnabled": false
     }
   ]
 }