Security audit

clean-web-fetch

Security checks for vulnerabilities and agentic risk

Overview

This is a web-page extraction helper with purpose-aligned behavior, but its package only includes instructions and references helper scripts that are not included.

Install only if you understand that this skill is meant to fetch external webpages and may expose the requested URL to remote servers. Do not use it with private, internal, authenticated, localhost, or token-containing URLs, and review or provide the missing helper script yourself before allowing the agent to run it or install Python dependencies.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

Confidence: 85% confidence
Finding: The skill instructs the agent to prioritize this tool for broadly defined tasks like fetching webpage content, article extraction, and converting pages to markdown. That broad activation scope can cause the agent to send arbitrary user-supplied URLs to an external target more often than necessary, increasing the chance of unintended network access, SSRF-like misuse in connected environments, or privacy-sensitive fetches.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill describes how to fetch arbitrary URLs but does not warn users that using it transmits the provided target over the network to third-party hosts. This can lead to inadvertent disclosure of sensitive URLs, tokens embedded in query strings, internal endpoints, or user intent metadata when the agent fetches content on the user's behalf.

Static analysis

No suspicious patterns detected.