urlgo

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed browser-control skill with powerful but purpose-aligned capabilities, though users should supervise it carefully.

Install only if you want an agent to control a browser. Verify where the external `urlgo` executable comes from, supervise click/type/eval actions, avoid running it on sensitive authenticated pages unless intended, and close the background browser when finished.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger list contains very broad, everyday terms such as “浏览器”, “网页”, “http://”, and “https://”, which can cause the skill to auto-activate in many unrelated conversations. Because this skill can open webpages, execute JS, click, type, and take snapshots, unintended invocation increases the chance of navigation to attacker-controlled content or execution of risky browser actions without clear user intent.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The Keywords section repeats broad activation phrases including URL prefixes and generic web-related terms, reinforcing accidental routing of ordinary user requests into this powerful browser-control skill. In context, that is more dangerous than a benign informational skill because this one can interact with live pages and execute arbitrary JavaScript via CDP.

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal