Neomano TTS (ElevenLabs)

Text-to-speech (TTS) via ElevenLabs. Use when the user asks to reply with voice/audio, generate a spoken version of some text, or asks for “voz”, “nota de vo...

Security Scan

The skill is internally consistent: it implements ElevenLabs TTS, only requires python3 and an ElevenLabs API key, and the included script and SKILL.md align with that purpose.

Purpose & Capability

Name/description (ElevenLabs TTS) match the requested env var (ELEVENLABS_API_KEY), required binary (python3), and the included script which POSTS text to the ElevenLabs text-to-speech endpoint.

Instruction Scope

SKILL.md instructs storing the API key in ~/.openclaw/.env, running the provided script with text and output path, and confirming the output file. The instructions do not request unrelated files, credentials, or external endpoints beyond ElevenLabs. They do instruct storing an API key in an env file (standard practice) — avoid committing that file.

Install Mechanism

No install spec (instruction-only plus a small script). No downloads or external install steps; risk is low because nothing arbitrary will be fetched or written during installation.

Credentials

Only ELEVENLABS_API_KEY (required) and ELEVENLABS_VOICE_ID (optional) are requested, which are appropriate and proportionate for calling ElevenLabs TTS. No unrelated secrets or config paths are requested.

Persistence & Privilege

Skill is not always-enabled and does not request elevated privileges or modify other skills or system-wide settings. It writes output audio files only to the path provided when invoked.

Assessment

This skill appears to do exactly what it says. Before installing: (1) only provide an ElevenLabs API key scoped to TTS and avoid committing ~/.openclaw/.env; (2) be aware that any text you send is transmitted to ElevenLabs (avoid sending secrets/PII in the text); (3) verify billing/usage on your ElevenLabs account because TTS calls may incur costs; and (4) confirm you have permission to synthesize any voice/identity you use (legal/consent considerations).

Human setup (one-time)

1. Create or edit ~/.openclaw/.env on the machine running OpenClaw.
2. Add your credentials (do not commit these):

ELEVENLABS_API_KEY=your_elevenlabs_key
# Optional (recommended): default voice
ELEVENLABS_VOICE_ID=your_voice_id

3. Restart the OpenClaw gateway (so the service environment picks up changes).

Inputs to collect (if missing)

• Text to speak.
• Optional: voiceId (ElevenLabs voice id). If not provided, use the default.
• Optional: output format (mp3 default).

Requirements (credentials)

This skill does not embed secrets. Credentials must be provided at runtime:

• ELEVENLABS_API_KEY (required)
  • Recommended: put it in ~/.openclaw/.env on the machine running the gateway.
• ELEVENLABS_VOICE_ID (optional but recommended)
  • If omitted, you must pass --voice-id when calling the script.

Example user prompts (to trigger this skill)

• "Reply with a voice note saying: …"
• "Generate audio (ElevenLabs) for: …"
• "Envíame una nota de voz que diga: …"

Workflow

1. Choose output path under the workspace, e.g. ./media/elevenlabs-tts/<timestamp>.mp3.
2. Run:

python3 {baseDir}/scripts/tts.py --text "..." --out "/abs/path/to/file.mp3" [--voice-id "..."]

3. Confirm with a short message including the output file path.

Defaults

• No voice is hardcoded. Set ELEVENLABS_VOICE_ID (recommended) or pass --voice-id.
• Model: eleven_multilingual_v2.