ClawHub
Back to skill

Security audit

thesis-review

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed thesis-review helper that reads thesis content, researches context, and creates a Word review document as its stated deliverable.

Before installing, expect the skill to analyze thesis documents, use web search for literature comparison, and create a Word review file locally. Confirm the requested output path and filename when using it, especially for repeated reviews of the same author on the same date.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (3)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly requires creating and saving a .docx file to the working directory as part of normal execution, but it does not instruct the agent to obtain explicit user consent, confirm the save location, or disclose that a file will be written. Unprompted file creation is a real safety issue because it causes side effects on the host environment and may overwrite files, leak sensitive content into persistent storage, or violate user expectations.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger phrases are broad terms commonly used in ordinary academic assistance requests, so the skill may be invoked when a user did not explicitly ask for this specialized thesis-review workflow. That creates ambiguous routing and can cause the agent to apply a domain-specific reviewing framework, generate formal evaluation content, or steer output in ways the user did not intend.

Natural-Language Policy Violations

Medium
Confidence
76% confidence
Finding
The description specifies Chinese-language behavior and a China-specific academic review workflow without indicating that language choice depends on user preference. While not a direct security flaw, this can cause unintended behavior, reduce user control, and increase the chance that the skill overrides the broader system or user context when invoked.

Static analysis

No suspicious patterns detected.