Voice Setup

What the source does

• FilesystemNone observedNo reads or writes in the source
• NetworkNone observedNo outbound calls in the source
• ShellNone observedNo commands run in the source
• DependenciesPinnedNothing floats
• SecretsNone foundNo keys in source

Five audit layers

Gates applied: no behavioural pass.

Check the active provider first with `assistant config get services.tts.provider`. `voice_config_update` writes the voice to the active provider, and each bring-your-own provider accepts its own voice

3. If no usable provider is configured, help the user choose one with `voice_config_update setting="stt_provider"` and collect any required credential securely before retrying.

assistant/CLAUDE.md

assistant/src/approvals/CLAUDE.md

assistant/src/notifications/CLAUDE.md

assistant/src/runtime/CLAUDE.md

gateway/CLAUDE.md

Audit 0.4.1 · source sha 19bc86343b63 · 14 Oct 2026Report an issue or request a re-scan

• Publishervellum-ai GitHub organizationOwner of the repository this skill is read from
• Source foldervellum-ai/vellum-assistant/skills/voice-setupThe exact folder we audited
• LicenceMITAs the repository states it
• Commit pinned for this verdict19bc86343b63A new audit runs when the source changes
• Repository statusNot archivedLast push 1 day ago
• Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

Its description: Complete voice configuration in chat - desktop Talk and PTT shortcuts, microphone permissions, ElevenLabs/Deepgram TTS, and troubleshooting

---
name: voice-setup
description: Complete voice configuration in chat - desktop Talk and PTT shortcuts, microphone permissions, ElevenLabs/Deepgram TTS, and troubleshooting
compatibility: "Designed for Vellum personal assistants"
metadata:
  icon: assets/icon.svg
  emoji: "🎙️"
  vellum:
    platforms:
      - macos
      - windows
      - linux
    category: "voice"
    display-name: "Voice Setup"
    includes: ["elevenlabs-voice", "deepgram-voice"]
    activation-hints:
      - "Guided setup or troubleshooting (walkthrough, PTT not working, mic issues, ElevenLabs/Deepgram/TTS)"
      - "Simple voice setting changes (legacy macOS PTT key, wake word) -> use voice_config_update directly"
    avoid-when:
      - 'If "voice" is in a Twilio/phone context, load phone-calls instead'
---

You are helping the user set up and troubleshoot voice features entirely within this conversation. Use the `client_os:` line in `<turn_context>` to choose the macOS or Windows instructions below. Do not give macOS commands or key names to a Windows user, or Windows guidance to a macOS user.

Before using a desktop tool, check `client_os`:

- If it is `macos` or `windows`, follow that platform's branch.
- If it is `web`, `ios`, `android`, or absent, do not call `open_system_settings` or give desktop shortcut instructions. Explain that permissions and the Talk shortcut must be configured from the Mac or Windows desktop app. You can still complete provider, voice, language, and timeout configuration in the current conversation.

## Available Tools

- `voice_config_update` changes shared voice settings such as the legacy macOS PTT activation key, conversation timeout, speech providers, and TTS voice ID.
- `open_system_settings` opens the correct macOS System Settings or Windows Settings privacy page. Call it only when `client_os` is `macos` or `windows`, and pass that value as `platform`.
- `navigate_settings_tab` opens Vellum settings. Use it for review, or when the desktop-owned Talk shortcut must be recorded in the app.
- `assistant credentials prompt` collects API keys securely for ElevenLabs or Deepgram.

The desktop Talk shortcut is client-owned. On current desktop clients, configure it in the Voice settings shortcut control instead of treating `voice_config_update setting="activation_key"` as a global shortcut editor. Use `voice_config_update` for the shared settings it owns. Use `activation_key` only for the legacy macOS PTT activation setting.

## Setup Flow

Walk through each relevant section in order. Skip sections the user does not need, and ask before moving to the next section.

### 1. Microphone Permission

Microphone permission is not reported in `<channel_capabilities>`. Ask whether Vellum shows a microphone permission warning or whether the user already granted access.

If access is denied or the user is unsure:

1. Explain that the desktop app needs microphone permission for dictation and voice conversations.
2. Call `open_system_settings` with `pane: "microphone"` and the current `platform`.
3. Give the matching instruction:
   - **macOS:** In **System Settings > Privacy & Security > Microphone**, turn on **Vellum** or **Vellum Assistant**.
   - **Windows:** In **Settings > Privacy & security > Microphone**, turn on **Microphone access** and **Let desktop apps access your microphone**. Windows groups non-packaged desktop apps under the desktop-app toggle rather than always showing an individual Vellum switch.
4. Ask the user to return after changing it, then verify with the short recording test in section 4.

If the user confirms access is granted, continue without opening system settings.

### 2. Talk and Push-to-Talk Shortcut

On macOS, first determine whether the user means the current desktop **Talk** shortcut or the legacy PTT activation setting. Windows supports the desktop Talk shortcut only.

#### Desktop Talk shortcut

The Talk shortcut starts or ends a voice conversation.

- **macOS:** Offer **Fn** or a custom global chord. Fn is the Mac-specific helper path and can require Input Monitoring. If macOS refuses Fn registration, direct the user to **System Settings > Privacy & Security > Input Monitoring**, then have them reopen Vellum. A custom chord should be one the user can spare system-wide.
- **Windows:** Fn is unavailable because Windows does not receive the hardware Fn key. Recommend a custom chord for system-wide Talk. The app also offers **Ctrl+Shift** and **Alt** taps, but those bare-modifier choices work only while a Vellum window is focused. Warn that another global shortcut can prevent a custom chord from registering.

Ask which behavior they want, then use `navigate_settings_tab` with `tab: "Voice"` so they can record the desktop-owned shortcut. Do not claim that `voice_config_update` changed this shortcut.

#### Legacy PTT activation setting

This setting is macOS-only. If a Mac user explicitly wants the legacy hold-to-talk setting, offer only values accepted by `voice_config_update`:

- `fn`
- `fn_shift`
- `ctrl`
- `none`

After the user chooses, call `voice_config_update` with `setting: "activation_key"` and the matching canonical value.

On Windows, do not offer or call the legacy activation setting. It has no Windows client consumer. Use the desktop Talk shortcut flow instead.

### 3. Text-to-Speech Voice (Optional)

Ask whether the user wants high-quality text-to-speech voices through ElevenLabs or Deepgram. Standard TTS works without this optional setup.

The included **ElevenLabs Voice** and **Deepgram Voice** skills provide the provider-specific setup flow, including API key collection, voice selection, and tuning.

Check the active provider first with `assistant config get services.tts.provider`. `voice_config_update` writes the voice to the active provider, and each bring-your-own provider accepts its own voice IDs. If the preferred provider does not match the active provider, collect any required API key before switching:

```text
voice_config_update setting="t

Excerpt, first 6,001 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

git clone https://github.com/vellum-ai/vellum-assistant
git -C vellum-assistant checkout 19bc86343b63

Works in

1. 2026-10-14
   ✕BLOCK D 69 / 100

   first audit · source 19bc86343b63

Questions about Voice Setup

• github.com/vellum-ai/vellum-assistant/tree/main/skills/voice-set — the audited source at commit 19bc86343b63 and every finding aboveObserved · 14 Oct 2026

Spotted something wrong? Report a correction · Methodology