Back to skill

Security audit

Trello Api

Security checks across malware telemetry and agentic risk

Overview

This skill advertises a Trello integration but its only working instructions inspect the local network instead of using Trello.

Review before installing. This does not show malware or persistence, but the advertised Trello purpose is not supported by the provided commands. Install only if you are comfortable with the skill suggesting local network inspection, or wait for a version that replaces those commands with real Trello API workflows and clear scope disclosures.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

High
Confidence
98% confidence
Finding
The skill is presented as a Trello API integration, but its actual quick-start commands only enumerate local interfaces, summarize TCP connections, and perform DNS lookups. This mismatch is dangerous because it can trick an agent or user into running unrelated host/network reconnaissance under the guise of harmless SaaS integration, expanding data exposure and violating least surprise.

Intent-Code Divergence

High
Confidence
97% confidence
Finding
The documentation claims the free tier 'works now' with real commands and no setup, but the only commands shown are unrelated system/network utilities. This deceptive framing increases the chance that agents will execute reconnaissance commands believing they are enabling Trello operations, causing unintended disclosure of local environment information.

Missing User Warnings

Low
Confidence
93% confidence
Finding
The quick-start instructs execution of network-related commands without disclosing that they inspect local interfaces, active connections, and DNS resolution. Even though these are common admin tools, running them without warning in a skill marketed for Trello can expose environment details and normalize unexpected system probing.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.