PDF Watermark

Security checks across malware telemetry and agentic risk

Overview

This skill appears to do what it claims: add text watermarks to user-selected PDFs, with no evidence of hidden data access or persistence.

Use this skill only with PDFs or URLs you intend the agent to process. Prefer trusted PDF sources, avoid internal or sensitive URLs, and choose an output filename that will not overwrite the original document. Install required Python dependencies from trusted sources.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill explicitly instructs the agent to fetch a PDF from an arbitrary URL with curl, but provides no warning, validation, or restrictions around network access or untrusted remote content. This can expose the environment to SSRF-style access to internal resources, retrieval of malicious files, or unsafe processing of attacker-controlled PDFs, especially because the downloaded file is then passed directly into a local script.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal