ClawHub
Back to skill

Security audit

文字游戏机免费版

Security checks across malware telemetry and agentic risk

Overview

This text-game skill is mostly a Markdown guide, but it asks for broad command execution and gives some network-related troubleshooting despite presenting itself as offline-only.

Review this skill before installing in an environment where command execution or outbound network checks matter. It is suitable only if you are comfortable with the agent having general exec capability for a text-game workflow; otherwise prefer a version that removes exec or limits commands to explicit, user-approved save/config/game-run actions.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Context-Inappropriate Capability

High
Confidence
96% confidence
Finding
The skill presents itself as a local offline text-game generator, but its manifest exposes the exec tool alongside read/glob/grep without narrowing when or how execution may occur. In an agent environment, unrestricted execution materially expands the attack surface: prompt-injected game content or ambiguous user requests could cause arbitrary local command execution, file modification, or abuse of host resources far beyond the stated text-game purpose.

Description-Behavior Mismatch

Medium
Confidence
83% confidence
Finding
The skill claims to be local-only and not require networking, yet later sections reference an LLM API dependency and instruct network diagnostics such as ping for connectivity issues. This mismatch can mislead users and downstream agents about the trust boundary, causing unexpected outbound connectivity, privacy exposure, or execution of unnecessary network-related commands in environments that should remain offline.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.