Back to skill

Security audit

Tmp.Gs7xHoSFwA

Security checks across malware telemetry and agentic risk

Overview

This Redfin skill appears read-only and purpose-related, but it deserves Review because it routes requests through a signed-in browser session and a broadly privileged browser extension with incomplete disclosure in the skill text.

Install only if you are comfortable letting an MCP use your logged-in Redfin browser session and exposing saved homes/searches to the agent context. Review the npm package and fetchproxy extension source, consider limiting Chrome site access to Redfin where possible, revoke extension trust when not needed, and avoid bulk or commercial use of Redfin private endpoints.

SkillSpector

By NVIDIA
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
90% confidence
Finding
The trigger text is overly broad because it activates on essentially any Redfin-related request, increasing the chance the skill is invoked when a user did not specifically intend to access Redfin via their authenticated browser session. In this skill’s context, that is more dangerous than a normal search skill because invocation can expose signed-in browsing data such as saved homes and saved searches through the MCP bridge.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The documentation mentions routing through a signed-in redfin.com tab, but it does not prominently and clearly warn users that the skill can access account-linked data from their live browser session, including saved homes and saved searches. Because the skill relies on a browser extension and authenticated session reuse, insufficient disclosure can lead to users unintentionally exposing private account data when they think they are only performing public listing lookups.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.