ClawHub
Back to skill

Security audit

Discord语音助手

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent Discord voice assistant, but it handles live third-party speech and external callbacks without enough consent and data-handling controls.

Review before installing. Use this only in Discord channels where participants are clearly told that speech may be recorded, transcribed, processed by an agent, and sent to selected STT/TTS providers. Prefer local providers where possible, restrict allowedUsers and channel scope, disable auto-join unless needed, and do not enable callback_url unless it is a trusted HTTPS endpoint with minimal transcript exposure. Store Discord and provider tokens in environment variables or a secret store.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

High
Confidence
95% confidence
Finding
The skill explicitly records voice, performs transcription, and may transmit audio/text to external STT/TTS providers, but it does not provide clear consent, notice, or privacy handling guidance. In a Discord voice context this is significant because it can capture third-party speech in real time, creating legal, privacy, and policy exposure if deployed without informed participant consent.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
Documenting a callback_url without warning that results may be posted to an external endpoint creates a risk of silent data exfiltration or accidental disclosure. Given this skill processes speech transcripts and operational metadata, sending results to arbitrary URLs can expose sensitive conversation content if users are not clearly warned and the destination is not validated.

Static analysis

No suspicious patterns detected.