draw-paper

Security checks across malware telemetry and agentic risk

Overview

This skill does what it claims: it reads paper figure prompts, sends them to Gemini through browser automation, and saves generated images locally.

Install only if you are comfortable with the fixed local paths and with sending extracted figure prompts to Gemini. Use simple intended filenames, avoid sensitive paper content in prompts, and check the output folder because repeated runs may overwrite figure files.

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 (1)

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill writes generated files to a hard-coded local directory on the host filesystem without user confirmation, path validation, or disclosure of side effects. Even though the write location is fixed rather than arbitrary, this still creates an unauthorized filesystem-write behavior that can overwrite prior outputs, leak sensitive derived content into predictable locations, and violate least-privilege expectations for an agent skill.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal