ClawHub
Back to skill

Security audit

Google Contacts

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Google Contacts integration that can read and change contacts through Maton's OAuth proxy, with no hidden installer or exfiltration behavior found.

Install only if you trust Maton as the OAuth proxy for your Google Contacts. Confirm the active Google connection before use, review every proposed create/update/delete carefully, and be especially cautious with group deletion options that can also delete member contacts.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
89% confidence
Finding
The manifest description narrows the scope to contacts and contact groups, but the documented API also exposes `otherContacts`, which can reveal relationship-derived address book data the user may not expect. This scope mismatch can cause over-broad access or misleading user consent because operators may believe the skill excludes that data class when it does not.

Intent-Code Divergence

Low
Confidence
79% confidence
Finding
The skill states that all write operations require explicit user approval, but the documented write/delete endpoints are ordinary examples with no built-in confirmation mechanism, guardrail, or invocation contract enforcing that requirement. In an agent setting, policy-only guidance is easy to bypass accidentally, leading to unauthorized contact creation, modification, or deletion.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The invocation guidance is broad enough to trigger on generic requests about people, search, or address-book operations without tight boundaries on when this skill should be preferred. In multi-tool agent environments, ambiguous routing can cause unnecessary access to sensitive contact data or use of the wrong connected account.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.