ClawHub
Back to skill

Security audit

Kaleido

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Kaleido connector helper that currently lists only read-only actions and uses the OOMOL `oo` CLI as expected.

Install only if you use Kaleido through OOMOL and are comfortable with the `oo` CLI accessing your connected Kaleido account. Treat any future write or destructive action as requiring explicit review of the target and payload before approval.

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 (2)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest and description claim the skill is for 'searching and reading data', but the body explicitly discusses support for `[write]` and `[destructive]` actions. This mismatch can mislead an orchestrating agent or user into invoking a skill under read-only assumptions, increasing the risk of unintended state-changing operations if future actions are added or exposed.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The instruction to use this skill for 'ANY Kaleido request' is overly broad and can cause agents to route all Kaleido-related tasks through this skill without considering least privilege or direct API alternatives. In combination with the documented support for state-changing actions, this broad trigger increases the chance of unnecessary or unsafe invocation.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.