Back to skill

Security audit

Doc Beautifier

Security checks for vulnerabilities and agentic risk

Overview

This skill locally reformats user-provided Word documents and does not show hidden access, persistence, network use, or credential handling.

Install only if you want an opinionated Word-document formatter. It may apply Chinese-oriented fonts, A4 page settings, and centered page numbers, so use it on a copy and review the output before replacing an original document.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The manifest description lists generic activators such as "beautify," "format," "make a document look professional," and "clean up a document's formatting." These phrases are broad natural-language requests that may overlap with ordinary user speech and the description does not provide exclusion conditions or tighter activation constraints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The description promises a "professional general template," but the documented defaults are locale-specific, including 微软雅黑, 宋体, and A4 page layout. Because the skill presents these as general defaults rather than an opt-in or region-specific mode, it may impose a specific locale/style convention on users without choice.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The top-level documentation says the script preserves all original text content, implying formatting-only changes. However, lines L373-L377 and L335-L351 add new footer content containing page-number fields, which changes document content beyond pure reformatting.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

Across the template definitions, the script forces locale-specific fonts such as 微软雅黑 and 宋体 for titles, headings, and body text. This is a natural-language locale policy concern because the formatting behavior is tied to a specific language environment without user opt-in or a documented justification that the tool is China-specific.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The manifest describes reading a source .docx, detecting structure, and applying formatting such as fonts, spacing, indentation, and page layout. Adding page-number fields in footers is a content augmentation step not described in that scope, so actual behavior exceeds the stated beautification-only semantics.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.