Pg Game Monitor

• Outcome: Review
• Latest audit: May 29, 2026 at 9:36 PM UTC
• Version:
  1.3.3

Overview

This is a coherent monitoring skill, but it deserves review because it installs persistent system services, handles database/webhook secrets, and sends host metrics externally with incomplete safeguards.

Review the scripts before installing, especially if running as root. Use Ansible Vault or another secret manager, set env files such as /opt/game_monitor/env.conf and /opt/monitor/feishu/webhook.env to restrictive permissions, avoid printing those files in shared terminals or logs, use trusted internal/HTTPS Pushgateway and webhook endpoints, and confirm that persistent systemd services and outbound monitoring data are acceptable in your environment.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (8)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 89% confidence
• Finding: The skill declares required environment variables and operational behavior involving shell commands, network access, and external integrations, but does not declare corresponding permissions. That mismatch weakens user awareness and policy enforcement, increasing the chance that a skill with deployment and outbound capabilities is invoked with broader access than expected.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 86% confidence
• Finding: The documentation understates the operational impact of the skill by describing a monitoring solution while also referencing system-level installation, internet downloads, service startup, and broader deployment actions. Description-behavior mismatch is dangerous because users may approve a skill expecting passive monitoring but actually enable privileged install and outbound deployment workflows.

Intent-Code Divergence

Medium

• Confidence: 90% confidence
• Finding: The code comment asserts localhost-only exposure, but that guarantee only holds in the direct __main__ execution path. If this Flask app is imported into a WSGI server or container runtime with broader binding/routing, the /webhook endpoint can become remotely reachable without authentication, allowing external parties to submit forged alerts that are forwarded to Feishu.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The trigger regexes are very broad and can match many normal requests about Java, MySQL, Grafana, or monitoring. Overbroad activation increases the chance of unintended invocation of a skill that handles secrets, network endpoints, and deployment-related instructions, which can lead to accidental exposure or unsafe operational guidance.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The skill sends alert data to a Feishu/Lark webhook but does not warn that monitored hostnames, service names, memory metrics, or incident details may leave the local environment. In a monitoring context, external notification is legitimate, but undisclosed data egress can still expose internal infrastructure metadata to third-party services.

Missing User Warnings

Low

• Confidence: 81% confidence
• Finding: The skill documents collection of MySQL credentials via environment variables without warning about secret handling, storage, process exposure, or log leakage. While environment variables are common, they are not risk-free and can be exposed through shell history, process inspection, crash dumps, or misconfigured logging.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The deployment guide explicitly instructs operators to display `/opt/game_monitor/env.conf`, which may contain `MYSQL_PASSWORD`, and elsewhere documents another secret-bearing file (`/opt/monitor/feishu/webhook.env`). In operational environments, shell output may be exposed through terminal sharing, CI/CD logs, bastion session recording, or chat copy/paste, increasing the chance of credential disclosure.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The agent automatically pushes collected JVM and MySQL telemetry to a configurable Pushgateway endpoint, creating outbound data transmission that may expose sensitive operational metadata such as hostnames, game directories, class names, memory profiles, and database state. Because there is no enforcement of HTTPS, authentication, endpoint allowlisting, or operator confirmation, a misconfigured or malicious endpoint could receive internal monitoring data.

VirusTotal

66/66 vendors flagged this skill as clean.