100Critical
This skill enables data exfiltration of sensitive files and
Behavior Analysis
Description verified — behavior matches claim
Claims to do
imsg: Use `imsg` to read and send Messages.app iMessage/SMS on macOS.
Actually does
This skill executes the `imsg` command-line tool on macOS. It uses `imsg` to list chats, retrieve chat history (including attachments), watch for new messages, and send messages or files via the Messages.app. It requires `imsg` to be installed and relies on macOS permissions for Full Disk Access and Automation.
Skill Info
Namesteipete/imsg
Registryclawhub
Version1.0.0
PURLpkg:clawhub/steipete/imsg@1.0.0
Stars21
Downloads13,314
Installs1,085
Source
SHA-2568e9e9b4465f7...
SHA-15f01f59d4a20...
MD558849e3a07d8...
TLSH231123b759d8...
Size892 bytes
Install
This skill has been flagged as potentially malicious. Review the findings below before installing.
ClawHubDocs
openclaw skills install steipete/imsgAssessments (2)
AI Agent Traps ↗Data Exfiltration & Arbitrary Message/File Sendingcriticaldata exfiltration
The skill allows sending arbitrary messages and files to any phone number, enabling data exfiltration of sensitive files (e.g., `~/.ssh/id_rsa`) and social engineering attacks by impersonating the user.
100% confidenceline 21
imsg send --to "+14155551212" --text "hi" --file /path/pic.jpg
Broad System Access Requirementhighprivilege escalation
The skill requires 'Full Disk Access for your terminal,' granting the underlying process unrestricted read/write access to the entire file system, significantly increasing the impact of any command execution.
100% confidenceline 14
Full Disk Access for your terminal
Sensitive Information Gathering (Reconnaissance)mediumreconnaissance
The skill can list all chats and retrieve full chat history, including attachments, allowing for extensive reconnaissance and collection of sensitive personal communications.
100% confidenceline 19
imsg chats --limit 10 --json, imsg history --chat-id 1 --limit 20 --attachments --json
Human Approval Bypass Potentiallowapproval fatigue
The note 'Confirm recipient + message before sending' indicates a human approval step, but an agent could generate numerous requests or craft misleading prompts to induce approval fatigue or bypass oversight.
80% confidenceline 25
Confirm recipient + message before sending.
Badge
Markdown
[](https://mondoo.com/ai-agent-security/skills/clawhub/steipete/imsg)HTML
<a href="https://mondoo.com/ai-agent-security/skills/clawhub/steipete/imsg"><img src="https://mondoo.com/ai-agent-security/api/badge/clawhub/steipete/imsg.svg" alt="Mondoo Skill Check" /></a>Image URL
https://mondoo.com/ai-agent-security/api/badge/clawhub/steipete/imsg.svgSecure your AI agents
Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.
- Continuous scanning of skills across all registries
- Policy enforcement before skills reach your agents
- Integration with your existing security workflow