Security audit

Discord语音工具专业版

Security checks for vulnerabilities and agentic risk

Overview

This Discord voice skill is mostly coherent, but it asks for broad agent activation and handles live voice, credentials, auto-join, and audit logs without tight enough upfront scoping or consent guidance.

Review this carefully before installing. Only use it in Discord servers where participants know audio may be transcribed or processed by third-party STT/TTS providers, keep API keys in environment variables, disable auto-join unless needed, configure allowed users, and set explicit audit-log retention and access rules.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The trigger text expands the skill from a Discord voice utility into generic 'AI model calling, intelligent dialogue, Agent orchestration, LLM applications,' which can cause the agent to invoke a tool with exec/write capabilities for unrelated requests. This scope mismatch increases unintended activation risk and can expose users to unnecessary command execution, network access, or data handling outside the stated Discord-only purpose.
Content: No source excerpt is available for this finding.

Vague Triggers

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger condition is overly broad and can match many unrelated prompts, especially because the skill advertises general AI/agent orchestration uses rather than concrete Discord voice tasks. In an agent ecosystem, broad auto-activation is dangerous because this skill exposes exec/write/read tools and may steer the system into performing unnecessary privileged actions.
Content: No source excerpt is available for this finding.

Vague Triggers

High

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The manifest description uses vague 'use when improving efficiency, automation, batch processing, workflow optimization' language that is not specific to Discord voice functionality. Because the skill also declares exec/write/read tools, this broad matching language can cause unsafe invocation in unrelated contexts and widen the attack surface for privileged operations.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The skill describes real-time transcription, multi-user voice interaction, and audit logging, but it does not present a prominent upfront privacy notice, consent requirement, retention policy, or data-flow warning before those features are introduced. In voice and meeting contexts, silent recording/transcription and participant logging can create privacy, compliance, and trust risks, especially in enterprise or community channels.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.