Security audit

gpx-kml-visualizer

Security checks across malware telemetry and agentic risk

Overview

This GPS route visualizer is mostly purpose-aligned, but it bundles a real-looking private route and local personal file path that users should review before installing.

Install only if you are comfortable with a reviewable GPS-processing tool that includes an unnecessary real-looking route sample. Remove the bundled track_tianmu.json before sharing or using the skill, avoid opening interactive HTML for sensitive routes unless external tile requests are acceptable, and use trusted GPX/KML inputs because generated outputs can preserve filenames and precise coordinates.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 92% confidence
Finding: The skill instructs the agent to read user-supplied GPX/KML files and write derived JSON/JPG/HTML outputs, but it does not declare any permissions. Undeclared file access weakens security review and policy enforcement because a caller may not realize the skill needs local filesystem capabilities, increasing the chance of unauthorized file reads or writes if the skill is invoked in a broader context.

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The generated HTML map loads map tiles from external providers (Esri and OpenTopoMap), which causes the output artifact to make network requests when opened. This can leak usage metadata such as viewer IP address, access time, and approximate area of interest, and it introduces dependency on third-party infrastructure that is not necessary for simple local visualization. In a GPS track visualizer, this is more sensitive because route maps may reveal personal location history.

Vague Triggers

Medium

Confidence: 81% confidence
Finding: The trigger language is broad enough to match generic mapping, charting, route planning, or diagram-generation requests that are not actually about GPX/KML processing. Overbroad activation can cause the wrong skill to run, leading to unnecessary file handling, creation of HTML artifacts, or confusion about what data is being processed, which expands the attack surface and can bypass user intent boundaries.

Missing User Warnings

Low

Confidence: 88% confidence
Finding: The skill generates interactive HTML that relies on external map tile requests, but this network behavior is only briefly noted and not clearly surfaced as a privacy/security warning. Opening the generated HTML can disclose route location, access patterns, IP address, and timing metadata to third-party tile providers, which is especially sensitive for GPS tracks representing personal movement.

Natural-Language Policy Violations

Low

Confidence: 96% confidence
Finding: The JSON embeds a full Windows absolute path pointing into a personal-data directory and chat/file-transfer area, which leaks local filesystem structure, user-associated identifiers, locale context, and potentially sensitive provenance metadata. In a GPX/KML visualization skill, this is unnecessary for route rendering and increases privacy risk if logs, outputs, or bundled assets are shared with other users or systems.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.