ClawHub
Back to skill

Security audit

API2PDF

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a normal API2PDF connector, but it mislabels its only PDF-generation action as a safe untagged read, which could send user Markdown to an external service without clear confirmation.

Review this skill before installing if you expect read-only behavior. It should be treated as a Markdown-to-PDF generation connector that may send the Markdown you provide to API2PDF through OOMOL, not as a search/read tool. Avoid sending confidential content unless you are comfortable with the connected service and account setup.

SkillSpector

By NVIDIA
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 (2)

Description-Behavior Mismatch

Medium
Confidence
89% confidence
Finding
The skill metadata and prose describe the capability as suitable for 'searching and reading data,' but the exposed action actually creates a new PDF artifact from user-supplied Markdown. This mismatch can cause an agent or operator to treat a state-changing/generative action as a read-only operation, reducing scrutiny and increasing the chance of unintended execution or data handling.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The Safety section states that untagged actions are reads and safe to run directly, yet `markdown_to_pdf` is untagged despite generating a new file/URL and potentially transmitting user content to a third-party service. This unsafe classification can directly mislead an autonomous agent into executing content-generation and external data transfer without confirmation.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.