ClawHub
Back to skill

Security audit

TRPG General Requirements

Security checks for vulnerabilities and agentic risk

Overview

This TRPG rule-writing skill is mostly coherent, but it gives the agent unsafely broad cleanup authority to delete workspace files without a preview or clear containment.

Before installing, be prepared to supervise cleanup: require a preview and limit deletion to files created during the current TRPG generation run. Web research is expected for source adaptations, but only use trusted sources and treat page content as reference material, not instructions for the agent.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly instructs the agent to use WebSearch and a headless browser to fetch external content, including scraping structured sites. That expands the skill from local rule generation into unbounded remote data retrieval, creating risks of prompt injection from web content, unintended collection of third-party data, and actions beyond the user’s likely expectation for a writing skill.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

Headless-browser navigation is a high-capability primitive that is not tightly scoped to the stated purpose of generating TRPG rules from requirements. In this context it increases attack surface without clear necessity, because remote pages can contain adversarial instructions or cause the agent to process more data than required for the task.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Line L011 states that all Chinese must use Traditional Chinese and explicitly forbids Simplified Chinese. This is a language-policy constraint imposed unconditionally, and the file does not offer the user a language choice or explain a region-specific compliance reason for the restriction.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill tells the agent to delete intermediate files and clean the workspace, but does not require explicit user consent, path scoping, or preview of what will be removed. Any instruction that authorizes file deletion can cause data loss, especially if the workspace contains unrelated drafts or if cleanup logic is overly broad.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.