Back to skill

Security audit

Realestate Advisor

Security checks across malware telemetry and agentic risk

Overview

This real-estate advisor skill is mostly purpose-aligned, but it includes automatic background maintenance that can modify or delete workspace memory/archive files without clear user control.

Review the heartbeat behavior before installing. If you do not want the skill to run scheduled cleanup or alter memory/archive files automatically, disable or edit HEARTBEAT.md. When using the skill, redact names, phone numbers, exact unit numbers, account identifiers, chats, and QR codes from screenshots before sharing them.

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 (4)

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The heartbeat tasks instruct automatic deletion of temporary logs and old archive files without any confirmation, backup, retention safeguards, or distinction between disposable and potentially important data. In a scheduled or gateway-triggered maintenance context, this can cause silent data loss, destroy audit trails, and remove information needed for troubleshooting or compliance, making the operational impact more serious than in a purely manual checklist.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill instructs users to provide listing-platform screenshots and relies on network-sourced housing data without clearly warning about personal or sensitive information that may appear in screenshots, such as names, phone numbers, addresses, chat history, or account details. In a real estate context, these materials can reveal highly sensitive financial and location data, increasing privacy and profiling risk if collected or retained without explicit minimization guidance.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The skill defines a very low-friction trigger where only a community name and area are enough to launch a full analysis workflow. This can cause unintended activation on ambiguous user messages and push the agent into collecting or inferring sensitive real-estate context without clear confirmation of user intent.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The automatic mode-selection logic relies on broad natural-language cues such as 'community + area' or general buying/swap intent, but it does not define strict precedence, exclusions, or ambiguity handling. This can misclassify user intent and produce advice or data processing in the wrong mode, increasing the chance of inappropriate analysis and over-collection of context.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.