Security audit

Audio Reply

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent local text-to-speech helper with disclosed URL-fetching and local audio playback behavior, plus explicit guardrails for public-only content.

Install this only if you want Claude Code to run local TTS commands and play audio. Use public, non-sensitive URLs only, avoid private or signed links, and be aware that broad phrases like "speak" or "say it" may trigger audio generation unexpectedly.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The documented trigger phrases include very broad terms like "speak", "say it", and "voice reply", which can easily appear in normal conversation and cause the skill to activate unintentionally. In this skill, unintended invocation is more concerning because activation can lead to local command execution for TTS and, in URL mode, network fetching of user-supplied public content.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: Telling users to "just use natural language" without precisely defining activation boundaries makes it unclear when the skill should run, increasing the chance of accidental execution. That ambiguity is risky here because the skill may fetch remote content or invoke local TTS tooling based on loosely matched user phrasing.

Vague Triggers

Medium

Confidence: 82% confidence
Finding: Using "etc." after listing trigger phrases leaves the trigger scope open-ended and undocumented, which can cause overly broad matching and unintended skill invocation. In the context of this skill, accidental activation could result in fetching external URLs or generating and playing audio when the user did not clearly request it.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger phrases "speak" and "say it" are very broad and can easily appear in ordinary conversation, increasing the chance of accidental activation. In this skill, unintended activation could cause unsolicited audio playback and, if paired with the URL-reading mode, could also lead to unexpected fetching of user-supplied public links.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The phrase "talk to me [topic/question]" is ambiguous and could match many harmless conversational requests not intended to invoke this skill. Because the skill produces spoken output and may suppress normal text output, accidental activation could disrupt the user experience or trigger unwanted actions without clear consent.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.