Back to skill

Security audit

pdf_to_word

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed SCNet PDF-to-Word wrapper that uploads a user-selected file for remote conversion, with clear privacy warnings and no hidden persistence or unrelated behavior.

Install only if you are comfortable sending the selected PDF to SCNet for remote processing. Do not use it for confidential, regulated, or business-sensitive documents unless you have separately reviewed SCNet's data handling terms and retention practices.

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill documentation indicates capabilities to read local files, access environment variables, invoke Python/shell, and make network requests, but it does not declare permissions explicitly. This weakens user and platform visibility into sensitive operations, especially because the skill uploads local documents to a third-party service and relies on an API key from the environment.

Description-Behavior Mismatch

High
Confidence
91% confidence
Finding
The documented `ocr_type` values include `IMAGE_TO_PPT`, which exceeds the skill's declared PDF/Image-to-Word purpose and could enable capability expansion beyond what users or platform reviewers expect. This mismatch is dangerous because it can facilitate unauthorized or undisclosed functionality, weakening trust boundaries and making policy enforcement harder.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill’s stated purpose is PDF/image to Word conversion, but the implementation uploads the user’s local file to a third-party SCNet service and only returns task metadata/results from that remote service. This is a real security-relevant mismatch because users may expect local/offline processing and may expose sensitive documents without clear consent from the manifest alone.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The skill sends user documents or externally hosted file URLs to a remote third-party conversion service, but the documentation does not warn about data transmission, storage, or privacy implications. This is risky because users may upload sensitive PDFs or images without understanding that the content leaves the local environment and may be retained or processed externally.

Static analysis

No suspicious patterns detected.