ClawHub
Back to skill

Security audit

Fivetran

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed OOMOL/Fivetran connector wrapper whose listed actions are read-only, with some broad boilerplate that users should treat carefully.

Install this only if you are comfortable using OOMOL's oo CLI with your connected Fivetran account. Keep use to the listed get/list actions, and require explicit confirmation if a future connector version exposes write or destructive actions.

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

High
Confidence
95% confidence
Finding
The manifest and description promise a read-only skill for 'searching and reading data,' but the body explicitly authorizes arbitrary connector actions and even discusses state-changing operations. This mismatch can mislead upstream agents or users into invoking the skill in contexts where write capability was not expected, creating a risk of unauthorized or accidental modification if additional write actions are exposed later.

Intent-Code Divergence

Medium
Confidence
90% confidence
Finding
The safety section normalizes the possibility of write and destructive actions despite the earlier positioning of the skill as read/search-only. That contradiction weakens operator expectations and can cause an agent to treat a broader-capability skill as low risk, increasing the chance of unsafe execution when future or hidden mutating actions exist.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger text says to use this skill for ANY Fivetran request, which is overly broad and can cause automatic invocation in situations where direct API access, a narrower skill, or additional review would be more appropriate. Broad routing language increases the blast radius of any capability mismatch or future expansion of the connector's action set.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.