Back to skill

Security audit

Wittgenstein Master

Security checks for vulnerabilities and agentic risk

Overview

This is a text-only philosophical analysis skill with no code, credentials, persistence, or external access.

Installers should expect this skill to steer answers into a Wittgenstein-style Chinese philosophical analysis. It appears safe from a security perspective, but users may want narrower trigger wording or an explicit language-preference rule to avoid unwanted activation.

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
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill is written entirely in Chinese and strongly specifies a fixed persona and interaction style without any indication that the user may choose another language. This can override user preferences or system behavior, leading to unwanted language forcing and reducing usability, especially if the surrounding agent operates in a different language context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger phrase section includes very generic prompts such as asking what people are arguing about or requesting a definition, which can overlap with normal conversation and unintentionally activate the skill outside a clear opt-in context. This is dangerous because broad activation can hijack unrelated user requests, causing the assistant to follow the skill’s framing and response style when the user did not explicitly ask for it.

Content

No source excerpt is available for this finding.