Paper to Slides

• Outcome: Pass
• Latest audit: May 29, 2026, 8:26 PM
• Version:
  1.0.0

Overview

This skill appears to do what it says: it turns user-provided papers into reports and HTML slides, with expected network and browser-preview behavior to be aware of.

Install if you are comfortable letting your agent read provided PDFs, download paper URLs, run poppler tools, and open a generated HTML presentation. Prefer local PDFs for confidential work, avoid untrusted PDF URLs, and review or disable remote fonts/localStorage if you need offline or privacy-sensitive use.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The README encourages users to provide arXiv links and PDF URLs but does not clearly disclose that doing so causes the agent to make outbound network requests and retrieve remote content. This is a real, low-severity security transparency issue because users may unknowingly trigger network access, expose metadata such as IP/request headers, or fetch untrusted remote documents in environments where egress is restricted or monitored.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill instructs the agent to fetch arbitrary remote PDF URLs with curl, including transformed arXiv URLs and other direct PDF links, without requiring user confirmation or warning about network access. This can cause unintended outbound requests, leak metadata such as IP/user-agent, and enable retrieval of attacker-controlled content from untrusted hosts.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The skill directs the agent to automatically open the generated HTML file via the system open command without explicit confirmation. Automatically launching generated HTML can trigger active browser content, external font fetches, and unintended local-user actions, especially when the HTML is derived from untrusted or attacker-influenced paper content.

VirusTotal

64/64 vendors flagged this skill as clean.