ClawHub
Back to skill

Security audit

humanize-zh-writing

Security checks for vulnerabilities and agentic risk

Overview

This is a text-rewriting skill that openly makes Chinese writing sound less AI-generated and has no code or system access, but users should avoid using it where AI-authorship disclosure is required.

Install this only if you want Chinese prose style editing aimed at making text sound more natural. Do not use it to bypass academic, workplace, publisher, or platform rules that require disclosure of AI assistance, and review rewritten text for unintended tone or meaning changes.

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
  • 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 (2)

Natural-Language Policy Violations

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill is explicitly designed to '去 AI 味', '消除 AI 写作痕迹', and make text appear as if written by a human, which facilitates concealment of AI authorship markers. Even without direct fraud instructions, this materially enables evasion of provenance detection, academic integrity controls, editorial disclosure requirements, or platform policies that depend on recognizing AI-assisted text.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger list includes broad, common-user phrases such as '帮我改得自然一点' and 'humanize', which can cause the skill to activate in benign editing contexts where the user did not intend this specialized behavior. That increases the chance of unintended rewriting toward AI-evasion goals and can route ordinary writing tasks into a more deceptive transformation flow.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.