Voice For Openclaw Publish

Security checks across malware telemetry and agentic risk

Overview

This voice-generation skill is coherent and purpose-aligned, but it uses external MiniMax/Telegram services and credentials that users should configure carefully.

This skill appears safe for its stated purpose if you want MiniMax-based text-to-speech. Before installing, understand that text you synthesize is sent to MiniMax, Telegram delivery can post generated audio to the configured chat, and credentials should be kept in a private .env file with only the services you actually need enabled.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

ASI03: Identity and Privilege Abuse

Low

What this means

Anyone installing the skill must trust it with the MiniMax API key, and Telegram credentials can allow posting voice messages to the configured chat.

Why it was flagged

The skill requires a MiniMax secret key and optionally uses Telegram bot credentials; these are expected for the stated TTS and message-sending purpose, but they are sensitive credentials.

Skill content

MINIMAX_API_KEY=your-minimax-secret-key-here
...
TELEGRAM_BOT_TOKEN=your-telegram-bot-token-here
TELEGRAM_TARGET=your-telegram-chat-id-here

Recommendation

Store the .env file securely, use only the credentials needed, restrict or rotate bot/API keys where possible, and avoid configuring Telegram credentials unless you want automatic voice-message delivery.

ASI07: Insecure Inter-Agent Communication

Low

What this means

Text submitted for speech generation is processed by MiniMax, and generated audio may also be sent to messaging services if those modes are used.

Why it was flagged

The script sends the text to be synthesized to MiniMax's external API. This is disclosed and central to the skill, but it means user-provided text leaves the local environment.

Skill content

API_BASE = "https://api.minimaxi.com"
...
data = {
        "model": model,
        "text": text,
...
resp = requests.post(url, headers=headers, json=data, timeout=120)

Recommendation

Do not synthesize confidential text unless you are comfortable sending it to MiniMax and, when enabled, to Telegram or Feishu.

ASI02: Tool Misuse and Exploitation

Low

What this means

A normal invocation can send a voice message to the configured Telegram chat, including groups if that chat ID is configured.

Why it was flagged

After generating audio, the default path posts it to Telegram if a target is available. This matches the skill's messaging purpose, but it is an external posting action rather than only local file generation.

Skill content

TARGET="${TARGET:-$DEFAULT_TARGET}"
...
RESPONSE=$(curl -s -X POST "https://api.telegram.org/bot${BOT_TOKEN}/sendVoice" \
    -F "chat_id=${TARGET}" \
    -F "voice=@${AUDIO_FILE}")

Recommendation

Confirm the Telegram target before use, prefer --generate-only when you only want a local audio file, and avoid configuring group/public chat targets unless intentional.

ASI04: Agentic Supply Chain Vulnerabilities

Info

What this means

The runtime behavior depends partly on whatever requests version is installed or later installed in the environment.

Why it was flagged

The skill depends on the Python requests package without a pinned version. This is common and purpose-aligned for API access, but it leaves dependency resolution to the user's environment.

Skill content

requires:
      bins: [python3, ffmpeg]
      env: [MINIMAX_API_KEY]
      pip: [requests]

Recommendation

Use a trusted Python environment and consider pinning or preinstalling a known-good requests version if reproducibility matters.