ClawHub
Back to skill

Security audit

TRPG General Requirements

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed TRPG rule-package generator that may create many project files, use web research when relevant, and clean up its own draft files.

Install this skill when you want an agent to produce a full TRPG rules package, not just a short rules draft. Use a clean project folder because it may create many files and remove intermediate drafts it believes are process artifacts. Be cautious when adapting existing works from the web and review generated files before publishing or sharing them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest describes a narrow rule-generation skill, but the body expands scope to broad project orchestration: creating many files, using external tools, and deleting workspace artifacts. This mismatch can cause the agent to perform actions the user did not reasonably expect, including destructive cleanup of files outside the intended deliverables.

Context-Inappropriate Capability

Low
Confidence
82% confidence
Finding
The skill instructs the agent to use WebSearch and a headless browser to retrieve external content even though the skill is presented as a general TRPG rule-writing aid. Unnecessary retrieval increases attack surface through prompt injection in fetched pages, untrusted content ingestion, and possible access to URLs the user did not explicitly authorize.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The trigger phrases are broad enough that the skill may activate for many generic content-generation requests, not just narrowly scoped TRPG requirements processing. Overbroad activation can unexpectedly impose hidden instructions, causing the agent to ignore user preferences or perform much larger tasks than intended.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.