ClawHub
Back to skill

Security audit

zotero-docx

Security checks for vulnerabilities and agentic risk

Overview

This skill is a focused local DOCX/Zotero rewriting tool with disclosed safeguards and no hidden persistence, credential use, or network behavior.

Install only if you are comfortable with a local Python tool modifying copies of your .docx files. Use dry_run first, keep original documents backed up, and do not pass allow_revisions=True unless you understand that inserted/deleted tracked-change text can be changed without creating new Word revision records.

SkillSpector

Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Intent-Code Divergence

High
Confidence
90% confidence
Finding
If the skill documentation tells users that text inside revision marks will not be rewritten, but the actual implementation still rewrites text inside w:ins regions, users may rely on a safety guarantee that does not exist. In this skill's context, that can corrupt tracked-change semantics, alter review history unexpectedly, and damage legal/academic document integrity while appearing compliant with the documented behavior.

Static analysis

No suspicious patterns detected.