discord-voice

ISSUES FOUND max severity HIGHSkillSpector Do not install 70 registry clawhub.ai last scanned 2026-02-06 1 scan

4 findings (2 high, 2 medium), mainly data exfiltration, resource abuse. Most recent scan 2026-02-06.

This Discord voice integration skill presents MULTIPLE HIGH-SEVERITY security concerns primarily related to credential management and data privacy. The skill requires multiple sensitive API keys (Discord bot token, OpenAI, Deepgram, ElevenLabs) and demonstrates poor security practices by showing hardcoded credentials in configuration examples. The skill processes highly sensitive biometric voice data from Discord users and transmits it to third-party AI services without documented privacy controls, consent mechanisms, or data retention policies. A critical referenced file (voice.py) is missing from the package, preventing complete security analysis. Additional concerns include unbounded resource consumption risks, missing security guidance for credential management, and lack of privacy compliance documentation. The skill requires immediate security improvements before deployment, particularly around credential handling, privacy controls, and resource limits.
Scanner's overall assessment (LLM analyzer)

Findings4

Critical0

High2

Medium2

Low0

Analyzersstatic behavioral llm trigger meta

Second scanner: NVIDIA SkillSpector

VerdictDo not install 70

Score70 / 100

Issues28

CategoriesSupply Chain Data Exfiltration Privilege Escalation

Version2.12.0 (static)

Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.

What this skill does

Real-time voice conversations in Discord voice channels with Claude AI

Package: 15 files, 125 KB
SKILL.md: 1076 words, 12+ sections
Code: TypeScript (8)

SKILL.md covers: Discord Voice Plugin for Clawdbot · Features · Requirements · Installation · 1. Install System Dependencies · Ubuntu/Debian · Fedora/RHEL · macOS · 2. Install via ClawdHub · 3. Configure in clawdbot.json · 4. Discord Bot Setup · Configuration

Findings

HIGH Hardcoded API Keys in Configuration Example Data Exfiltrationllm

The configuration example in SKILL.md demonstrates hardcoding API keys directly in the clawdbot.json file ('openai.apiKey': 'sk-...'). This practice exposes sensitive credentials in configuration files that may be committed to version control, shared, or accessed by unauthorized parties. While the documentation mentions environment variables as an alternative, the primary example shows direct key embedding, which users may follow.

"openai": {
  "apiKey": "sk-..."  // Or use OPENAI_API_KEY env var
}

Remediation: Remove the hardcoded API key example from the configuration. Update documentation to exclusively recommend environment variables (OPENAI_API_KEY, ELEVENLABS_API_KEY, DEEPGRAM_API_KEY) for credential management. Add a security warning against hardcoding credentials in configuration files.

HIGH Multiple Sensitive API Keys Required Without Security Guidance Data Exfiltrationllm

The skill requires multiple sensitive API keys (DISCORD_TOKEN, OPENAI_API_KEY, and optionally ELEVENLABS_API_KEY, DEEPGRAM_API_KEY) but provides minimal security guidance on credential management. Discord bot tokens provide full bot access, and AI API keys can incur significant costs if compromised. The configuration example shows credentials in plaintext JSON, increasing exposure risk.

requiredEnv: ['DISCORD_TOKEN', 'OPENAI_API_KEY']
optionalEnv: ['ELEVENLABS_API_KEY', 'DEEPGRAM_API_KEY']

Remediation: Add a dedicated security section covering: (1) Use environment variables exclusively for all credentials, (2) Never commit credentials to version control, (3) Use .env files with proper .gitignore configuration, (4) Implement credential rotation procedures, (5) Use least-privilege API keys where possible, (6) Monitor API usage for anomalies.

MEDIUM Unbounded Audio Recording Without Resource Limits Resource Abusellm

The skill implements voice recording with a configurable maxRecordingMs (default 30000ms/30 seconds), but there are no documented limits on concurrent recordings, total storage, or aggregate resource consumption. Multiple users could simultaneously trigger 30-second recordings, potentially exhausting memory, disk space, or API quotas. The auto-reconnect feature could also lead to resource exhaustion if reconnection logic fails to implement backoff.

"maxRecordingMs": 30000,
"Auto-reconnect": Automatic heartbeat monitoring and reconnection on disconnect

Remediation: Implement and document: (1) Maximum concurrent voice sessions per guild/bot, (2) Rate limiting for voice channel joins, (3) Exponential backoff for auto-reconnect attempts, (4) Maximum total recording storage limits, (5) Cleanup procedures for temporary audio files, (6) API quota monitoring and circuit breakers.

MEDIUM Audio Data Processing Without Privacy Controls Data Exfiltrationllm

The skill captures, transcribes, and processes voice audio from Discord users, then sends this data to third-party AI services (OpenAI, Deepgram, ElevenLabs). There is no documentation of: (1) Audio data retention policies, (2) User consent mechanisms, (3) Data deletion procedures, (4) Privacy implications of sending voice data to external APIs, (5) Compliance with privacy regulations (GDPR, CCPA). Voice data is highly sensitive biometric information.

Speech-to-Text: Whisper API (OpenAI) or Deepgram
Text-to-Speech: OpenAI TTS or ElevenLabs
Transcribed speech is routed through the Clawdbot agent

Remediation: Add comprehensive privacy documentation: (1) Explicit user consent requirements before voice recording, (2) Data retention and deletion policies, (3) Third-party data sharing disclosure, (4) Privacy policy requirements for bot operators, (5) Options to disable voice recording per user/guild, (6) Audit logging of voice data access, (7) Compliance guidance for GDPR/CCPA.

What is in the package

File	Kind	Size
`CHANGELOG.md`	doc	1.5 KB
`README.md`	doc	7.1 KB
`SKILL.md`	doc	8.1 KB
`clawdbot.plugin.json`	config	3.8 KB
`index.ts`	script	19 KB
`package-lock.json`	config	17 KB
`package.json`	config	966 B
`tsconfig.json`	config	360 B
`src/config.ts`	script	5.2 KB
`src/core-bridge.ts`	script	5.9 KB
`src/streaming-stt.ts`	script	11 KB
`src/streaming-tts.ts`	script	5.8 KB
`src/stt.ts`	script	4.7 KB
`src/tts.ts`	script	3.3 KB
`src/voice-connection.ts`	script	31 KB

External hosts referenced

registry.npmjs.org github.com api.openai.com api.elevenlabs.io api.deepgram.com keepachangelog.com semver.org

Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.

Environment variables referenced

OPENAI_API_KEY DEEPGRAM_API_KEY ELEVENLABS_API_KEY DISCORD_TOKEN

Package contents recorded 2026-10-22. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.

What we checked

Static analyzer. Pattern rules over SKILL.md and bundled files: known-bad commands, exfiltration URLs, secrets, obfuscation.
Behavioral analyzer. Checks what the skill asks the agent to do against what its description claims it does.
Llm analyzer. A language model reads the skill and reports intent-level issues the rules cannot express.
Trigger analyzer. Looks for instructions that fire on conditions the user did not ask for.
Meta analyzer. A second model pass that validates, deduplicates and re-scores the other analyzers' findings.

What we did not check

Runtime behaviour. The skill was not executed; findings come from reading its files.
Anything fetched at install or run time, including remote scripts, package dependencies, and URLs the skill tells the agent to open.
Content the analyzers cannot read: images, binaries, and some non-English text.
Known scanner bypass classes. See the methodology and limitations page.

Sources

Repository: https://clawhub.ai/skills/discord-voice
Raw scan result: discord-voice-scan.json
Badge for your README: ![skillscan.io](https://skillscan.io/badge/clawhub-ai-discord-voice.svg)
Request a re-scan