Claw Messenger

ReviewAudited by ClawScan on May 10, 2026.

Overview

The skill is coherent for messaging, but it asks the agent to install unreviewed plugin code, store a live messaging API key, and send messages through third-party infrastructure without clear outbound approval limits.

Before installing, make sure you trust the Claw Messenger provider and the external plugin package. Use a dedicated API key, limit who the agent may message, require approval before outbound messages or media, and understand that message content and routing data pass through the provider's servers.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Installing this gives unreviewed external code access to the messaging channel and API key.

Why it was flagged

The skill instructs installation of an external plugin, while the provided review set contains no plugin code or install spec. That unreviewed package would handle live messaging and credentials.

Skill content
openclaw plugins install @emotion-machine/claw-messenger
Recommendation

Install only if you trust the provider and package; prefer reviewed source, pinned versions, and a declared install spec before giving it credentials.

What this means

The agent or plugin could use the live key to send and receive messages and potentially consume paid quota.

Why it was flagged

The skill requires a live service credential and stores it in agent configuration, despite registry metadata declaring no primary credential or required config path.

Skill content
Ask your human for the API key from the dashboard (it starts with `cm_live_`) ... `apiKey` (required): The live API key from the dashboard.
Recommendation

Use a dedicated, revocable key; restrict routes where possible; rotate the key if exposed; and require the skill metadata to declare the credential and config file.

What this means

An agent could send unintended messages or media to people and incur usage charges.

Why it was flagged

The instructions authorize broad outbound messaging and media/group-chat actions, but do not require human approval per message or define outbound allowlists or spending limits.

Skill content
Once configured, use this channel for all iMessage, RCS, and SMS communication. ... You can send text messages and media ... and participate in group chats.
Recommendation

Require explicit human confirmation for outbound messages, configure recipient allowlists where possible, and monitor billing and message logs.

What this means

Private communications may be processed by the provider's infrastructure.

Why it was flagged

Message contents, media, phone numbers, and routing metadata pass through a third-party WebSocket service. This is disclosed and purpose-aligned, but it is a sensitive data boundary.

Skill content
plugin sends it over WebSocket -> Emotion Machine server routes it ... Someone texts the shared number -> Emotion Machine server looks up phone route -> forwards to your WebSocket connection
Recommendation

Review the provider's privacy, retention, and logging practices before using it for sensitive messages.