Activity Control Ui

• Outcome: Review
• Latest audit: May 29, 2026 at 9:29 PM UTC
• Version:
  1.1.0

Overview

This appears to be a dashboard skill, but its local server exposes control and file access in ways that need review before installation.

Install only if you are comfortable running a local server from this skill. Bind it to localhost, avoid exposing it through tunnels or LAN interfaces, require a strong unguessable token for any control action, and review or restrict the static file handler before use.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 93% confidence
• Finding: The skill is presented mainly as a visual dashboard, but the documented behavior includes running an HTTP/WebSocket server, exposing an API, accepting inbound client messages, and allowing a client-triggered compact action. Those behaviors expand the attack surface significantly: local services can be abused by other local processes, browser-origin requests, or unintended network exposure if bound insecurely.

Context-Inappropriate Capability

Medium

• Confidence: 97% confidence
• Finding: The HTTP handler uses req.url directly and joins it with the skill directory, then serves whatever file resolves there. This exposes any file under the skill directory to anyone who can reach the server, not just intended UI assets, which can leak source code, configuration, prompts, or other sensitive local files packaged with the skill; depending on path normalization behavior, it may also enable traversal outside the intended assets area.

VirusTotal

63/63 vendors flagged this skill as clean.