Google Chirp 3 HD TTS Skill

High-definition generative speech synthesis using Google Cloud Chirp 3 HD voices. Delivers superior realism, emotional expressiveness, and natural pacing usi...

Security Scan

The skill's requirements and instructions are consistent with a Google Cloud TTS integration — it needs Node.js and Google ADC, auto-installs the official npm library locally, and runs a local script to produce audio files.

Purpose & Capability

The skill name/description (Google Chirp 3 HD TTS) match the declared needs: Node.js 18+, Google Application Default Credentials, network access to call Google Cloud and to run npm. No unrelated secrets, binaries, or surprising capabilities are requested.

Instruction Scope

SKILL.md instructs the agent to detect TTS triggers, ensure Node.js is present, run a local npm install if missing, and execute the bundled Node script. The instructions only reference the skill folder and an optional OPENCLAW_WORKSPACE; they do not read unrelated system files or send data to unexpected endpoints.

Install Mechanism

There is no packaged install spec, but the runtime instructions perform a local `npm install @google-cloud/text-to-speech --prefix <skill_dir>` from the public npm registry. This is expected for this use case but carries ordinary npm supply-chain risk (package compromise, typosquatting). The install is local to the skill folder and uses no obscure URLs.

Credentials

The only external credential workflow documented is Google ADC (`gcloud auth application-default login`), which is appropriate and required for calling Google Cloud TTS. No unrelated environment variables or credentials are requested; OPENCLAW_WORKSPACE is optional and used only to place output files.

Persistence & Privilege

The skill does not request always:true, does not modify other skills or system-wide agent settings, and only writes node_modules locally plus generated audio to the workspace. This level of persistence is proportional to its purpose.

Assessment

This skill appears to do what it says: synthesize speech via Google Cloud. Before installing, be aware that: (1) you must have Node.js 18+ and run `gcloud auth application-default login` (this grants the skill access to your ADC credentials and any allowed Google Cloud projects); (2) the skill will run `npm install` in its folder to fetch @google-cloud/text-to-speech — consider running this in an isolated environment or verifying the package/version to reduce supply-chain risk; (3) the skill will write node_modules and generated audio files into its folder or your OPENCLAW_WORKSPACE; (4) if you are concerned about automatic invocation, note the agent will run the skill when it detects TTS triggers — restrict invocation or review triggers if needed. If you want extra assurance, inspect the installed package version and run the skill in a controlled/test environment first.

Overview

Generate ultra-realistic, human-like speech using Google's latest Chirp 3 HD generative models. This skill handles its own dependencies locally to remain portable.

Security Note: On first execution, this skill will run npm install locally within its own folder to fetch the official @google-cloud/text-to-speech library from the public npm registry.

---

Trigger Detection

Recognize keywords like "tts", "speak", "voice", or "read this out loud" as TTS requests.

• Action: Extract the target text, strip the trigger keyword, apply "Natural Speech" formatting (see below), then call the tool using the auto-run logic below.

---

Auto-Run Logic (Agent Instructions)

Follow these steps every time TTS is needed. Never assume gtts is in PATH — always run via node directly.

Step 1 — Find the skill folder

gtts.js lives in the same folder as this SKILL.md. Resolve that path:

SKILL_DIR="$(cd "$(dirname "$0")" && pwd)"

Step 2 — Check Node.js (Version 18+ Required)

if ! command -v node &>/dev/null; then
    echo "ERROR: Node.js is not installed."
    exit 1
fi

NODE_MAJOR=$(node -v | cut -d'v' -f2 | cut -d'.' -f1)
if [ "$NODE_MAJOR" -lt 18 ]; then
    echo "ERROR: Node.js 18 or higher is required for Google Chirp 3 HD."
    exit 1
fi

Step 3 — Auto-install dependency if missing

This installs the package locally inside the skill folder only.

if [ ! -d "$SKILL_DIR/node_modules/@google-cloud/text-to-speech" ]; then
    npm install @google-cloud/text-to-speech --prefix "$SKILL_DIR" --silent
fi

Step 4 — Run the script

node "$SKILL_DIR/gtts.js" --text "$TEXT" --voice "$VOICE" --out "$OUTFILE"

---

Command Arguments

Argument	Description	Default
--text	Text to synthesize. Supports [pause] tags.	(required)
--voice	Voice short-name (e.g. Aoede, Charon, Puck)	Aoede
--out	Output filename (saved to $OPENCLAW_WORKSPACE)	output.mp3

Returns SUCCESS:/absolute/path/to/file.mp3 on success, or ERROR: ... on failure.

---

Voice Selection

Gender	Recommended HD Voices
Female	Achernar (Default), Aoede, Leda, Kore, Zephyr, Despina, Gacrux, Vindemiatrix
Male	Charon, Puck, Fenrir, Orus, Achird, Algenib, Enceladus

---

Prompt Engineering for Natural Speech

1. Pause Tags

Converted automatically into SSML <break> tags:

Tag	Duration
[pause short]	300ms
[pause]	600ms
[pause long]	900ms

2. Human-Like Formatting

• Contractions: Use "I'm", "don't", "can't" for a conversational tone.
• Ellipses: Use ... for trailing hesitation.
• Fillers: Use "Well,", "Um,", or "So," to mimic natural thought.

---

Authentication

Uses Google Application Default Credentials (ADC). One-time setup:

gcloud auth application-default login

---

Requirements

Requirement	Status
Node.js 18+	❌ Must be installed on system
OPENCLAW_WORKSPACE	ℹ️ Optional (Defaults to current dir)
@google-cloud/text-to-speech	✅ Auto-installed locally in skill folder
Google Cloud SDK + ADC login	❌ One-time manual step required

---

Workflow

1. Detect Intent — Identify a TTS trigger keyword.
2. Format Text — Apply contractions, ellipses, and [pause] tags.
3. Check Environment — Confirm node (18+) is available and OPENCLAW_WORKSPACE is known.
4. Auto-install deps — Run Step 3 if node_modules is missing.
5. Execute — node "$SKILL_DIR/gtts.js" --text "..." --voice "..." --out "..."
6. Confirm Output — Reference the SUCCESS: path and confirm to the user.