bluebubbles

ReviewAudited by ClawScan on May 11, 2026.

Overview

The skill matches its iMessage purpose, but it grants broad authority to send or alter messages and attachments without clearly documenting confirmation or recipient limits.

Install only if you trust the BlueBubbles server and are comfortable letting the agent send or alter iMessages. Configure a trusted server URL, protect the server password, restrict allowed_senders to known contacts, and require explicit confirmation before sending messages or attachments.

Findings (3)

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

If the agent uses this capability incorrectly, it could send messages or files, react to messages, or alter conversations from the user's iMessage account.

Why it was flagged

The skill advertises the ability to send and modify iMessage conversations, including attachments, but the supplied instructions do not describe confirmation, recipient scoping, or review controls for these high-impact actions.

Skill content
Supports text messages, attachments, reactions, edits, unsend ... "tool": "bluebubbles_send" ... "target": "+15551234567"
Recommendation

Require explicit user confirmation before every send, edit, unsend, reaction, or attachment; restrict permitted recipients where possible.

What this means

Anyone or any agent process with this password may be able to access the BlueBubbles server and send messages through it.

Why it was flagged

The BlueBubbles server password is expected for this integration, but it grants access to the user's messaging server and is not reflected as a primary credential in the registry metadata.

Skill content
password = "your-server-password"
Recommendation

Store the password securely, rotate it if exposed, and ensure the registry/installation documentation clearly declares the required credential.

What this means

Unexpected contacts could potentially send messages into the channel, increasing the chance of unwanted prompts, data exposure, or accidental replies.

Why it was flagged

The example configuration allows all senders, which may broaden who can interact with the messaging channel if the gateway treats incoming iMessages as allowed inputs.

Skill content
allowed_senders = ["*"]  # Allow all senders, or list specific handles
Recommendation

Replace the wildcard with a small list of trusted handles and review how inbound BlueBubbles messages are routed to the agent.