Web Fetcher

Security checks across malware telemetry and agentic risk

Overview

This skill transparently fetches web pages through named services, with privacy cautions around sensitive URLs.

Install this for public web-page fetching and summarization. Do not use it with intranet links, signed links, password-reset URLs, URLs containing access tokens, or confidential pages unless you are comfortable sharing them with the named conversion services; attach browser tabs only when you intentionally want the agent to inspect that page.

SkillSpector

By NVIDIA

Vulnerability Patterns

MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
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)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 89% confidence
Finding: The skill explicitly instructs the agent to perform outbound web requests through multiple third-party fetching services and browser-based retrieval, yet no permissions are declared. Undeclared network capability weakens policy enforcement, auditability, and user awareness, and can enable unintended access to arbitrary external URLs or data exfiltration pathways through external services.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The script fetches arbitrary remote content and returns the full body verbatim in JSON without any disclosure, filtering, or trust boundary enforcement. In the context of an agent skill, this can expose users or downstream models to untrusted third-party content, including prompt-injection instructions, sensitive page data, or copyrighted/private material copied through proxy services.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal