Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent Discord voice-agent guide, but users should secure bot tokens and handle recorded voice/transcript data carefully.
Before installing, store Discord/OpenClaw tokens only in a secure environment or secret store, do not paste them into shared chats or commit them, and make sure Discord participants know the bot can capture, transcribe, and locally save voice turns and conversation history.
The skill instructs the operator to provide a Discord token and to auto-detect OpenClaw settings from the environment, but it does not include any warning or handling guidance for secrets. In an agent setting, this can normalize pasting credentials into chat, logs, or prompts, or encourage broad environment exposure without clarifying least-privilege or redaction expectations.
The documentation explicitly states that voice receive/capture uses per-turn saved audio, but it does not warn operators that user voice data may be recorded and retained on disk. In a Discord voice agent context, this can lead to unnoticed collection of sensitive audio, privacy violations, and noncompliant deployment if admins enable the feature without informed consent or retention controls.
The file lists on-disk locations for conversation history and captured voice turns, but it does not disclose that users' audio and chat-derived data will be persisted locally. This increases the risk that deployers unknowingly retain sensitive content, which can be exposed through host compromise, backups, shared systems, or simple operational oversight.
The README explicitly describes capturing speech, transcribing it, and routing replies through OpenClaw, but it does not warn operators that user voice data and transcripts may be collected, processed, retained, or transmitted to external/local services. In a Discord voice-agent context, this omission can lead to deployments that record or process users’ spoken content without informed consent, creating privacy, compliance, and trust risks.
The demo script presents the bot joining a Discord voice channel, listening, and replying, but it does not disclose that spoken audio will be captured and processed. In a voice-agent context, omission of a clear consent/privacy warning can mislead operators or participants, increasing the risk of non-consensual monitoring, privacy complaints, or policy non-compliance when the demo is reused as setup guidance or marketing material.
The wizard explicitly asks the user to provide DISCORD_TOKEN but gives no guidance that this is a secret, how it should be entered securely, or that it must not be echoed, logged, or committed. In an onboarding flow for an agent skill, this omission can lead to credential exposure through chat transcripts, screenshots, shell history, or config files, enabling takeover of the Discord bot account.
This markdown file instructs users to set DISCORD_TOKEN and DISCORD_VOICE_CHANNEL_ID, which relate to authentication and connecting a live voice agent to Discord. The release note does not include any warning about safeguarding the token, granting bot access, or the privacy impact of capturing/transcribing voice in a channel.
This markdown file instructs the user to set DISCORD_TOKEN but does not warn that the value is a secret credential or that it should not be logged, committed, or shared. Because markdown files should disclose behaviors affecting privacy or system integrity, the omission is a minor but real safety gap.
No suspicious patterns detected.
