ClawHub
Back to skill

Security audit

Ecologi

Security checks across malware telemetry and agentic risk

Overview

The skill mostly does what an Ecologi connector would do, but it under-discloses billable purchase actions and marks untagged actions as safe reads even though purchases are untagged.

Review this skill before installing. Use it only if you are comfortable with an OOMOL-connected Ecologi account being able to run purchase actions, and require explicit confirmation of the exact purchase payload and cost before any purchase_* action is executed.

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 (4)

Description-Behavior Mismatch

High
Confidence
97% confidence
Finding
The manifest advertises this skill as only for 'searching and reading data,' but the body exposes multiple purchase actions that can spend money and modify account state. That mismatch can cause an orchestrator or user to treat the skill as read-only and invoke it in contexts where state-changing operations would not otherwise be allowed or carefully confirmed.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The safety model says state-changing actions will be labeled with a '[write]' tag, but the purchase actions are listed without those tags. If an agent relies on those tags to decide when to ask for confirmation, the omission can lead to silent execution of purchase operations that spend funds or create unintended transactions.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The instruction to use this skill for 'ANY Ecologi request' is overly broad and can cause automatic routing of all Ecologi-related tasks into a skill that includes state-changing purchase operations. In context, that broad trigger increases the chance that a casual informational request or ambiguous prompt gets handled by a tool with spending capability.

Missing User Warnings

High
Confidence
97% confidence
Finding
The purchase actions are state-changing and potentially billable, yet they are not marked with the write/destructive tags that the safety section says govern confirmation behavior. In a tool-using agent environment, this creates a direct path to unauthorized purchases because the control intended to distinguish safe reads from dangerous writes is missing where it matters most.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.