ClawHub
Back to skill

Security audit

Google Analytics CLI

Security checks across malware telemetry and agentic risk

Overview

This GA4 reporting skill is mostly coherent, but it understates its authority while documenting commands that can create audience exports and expose sensitive analytics identity data.

Install only if you are comfortable granting the agent access to GA4 data. Use a dedicated least-privilege service account scoped to the intended properties, avoid broad default gcloud credentials, do not paste credential JSON into chat, and require explicit approval before running access-report or audience-export commands.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill explicitly labels the CLI as read-only, but later documents `audience-export-create`, which performs a state-changing operation in GA4. This mismatch can cause an agent or user to invoke modifying commands under the false assumption that the tool cannot change remote state, weakening safety controls and trust assumptions.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
This is a documentation integrity issue with security consequences: the skill claims the tool is read-only while instructing use of a create operation later in the file. In agent settings, inaccurate capability descriptions can bypass risk-based routing or cause unsafe automation because operators rely on the read-only label when deciding whether a tool is safe to invoke automatically.

Vague Triggers

Medium
Confidence
82% confidence
Finding
The trigger list contains very broad phrases such as general analytics terms, increasing the chance the skill is invoked for unrelated requests. Over-broad invocation can expose sensitive analytics/account metadata or cause the agent to prefer this skill when the user did not intend to access GA4 data.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The authentication section instructs the user to work with service account credentials and references sensitive analytics/access-report features, but does not warn about protecting credential files, avoiding secret disclosure, or handling potentially sensitive output such as accessor emails. In a skill context, missing guardrails around secrets and analytics identity data increases the risk of accidental credential exposure or privacy leakage.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.