Emily Web Fetch

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:16 PM UTC
• Version:
  1.0.5

Overview

This is a simple user-directed webpage fetcher with normal web-request privacy risks but no hidden install behavior, persistence, credential access, or destructive capability.

Install this only if you want the agent to make web requests on your behalf. Avoid fetching localhost, private network, cloud metadata, intranet, or otherwise sensitive URLs, and treat returned webpage text as untrusted content rather than instructions.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The skill enables outbound requests to arbitrary user-specified URLs but does not warn users or downstream agents that invoking it can contact external systems. This can lead to SSRF-like access attempts, unintended disclosure of IP/network metadata, or fetching sensitive internal endpoints if the runtime has broader network reach than the user expects.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: The skill hard-codes an `Accept-Language` header preferring `zh-CN`, which can alter content negotiation and disclose a language or regional preference the user did not choose. This creates an unnecessary privacy leak and may also bias fetched content in ways that are unexpected, especially in a generic web-fetching skill where users would reasonably expect neutral request behavior.

VirusTotal

65/65 vendors flagged this skill as clean.