ClawHub
Back to skill

Security audit

Cuttly

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly coherent for Cuttly use, but it under-discloses and under-controls a state-changing URL creation action while using a credentialed connector.

Review this skill before installing if you use Cuttly for business or public links. It should be treated as able to create short URLs with your connected account, not just read analytics, and you should require confirmation before any `shorten_url` action or first-time CLI/auth setup.

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
93% confidence
Finding
The manifest and description frame the skill as suitable for 'searching and reading data,' but the documented `shorten_url` action performs a state-changing operation. This mismatch can cause an agent or user to invoke the skill under read-only assumptions and accidentally create resources without the stricter confirmation normally required for write actions.

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The skill presents itself as read-oriented at the manifest level while later advertising a state-changing `shorten_url` capability. In agentic environments, this kind of semantic mismatch is dangerous because routing, approval, or policy decisions may rely on the manifest summary and therefore under-classify the skill's risk.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The phrase 'Use this skill for ANY Cuttly request' is an overly broad routing instruction that can force invocation in situations where direct API access, narrower tooling, or stricter approval flows would be more appropriate. Broad trigger language increases the chance of inappropriate automatic tool selection and expands the blast radius of any unsafe behavior in the skill.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.