Agent Browser Core.Conflict
Security checks across malware telemetry and agentic risk
Overview
This documentation-only skill describes browser automation with clear safety limits and no hidden code or destructive behavior.
Before installing, verify the publisher and the external agent-browser npm package, pin a reviewed version, and run it in a low-privilege dedicated environment. Require explicit approval before using eval, local file access, traffic interception, downloads, credential or session changes, or saved login state.
SkillSpector
By NVIDIA
Vulnerability Patterns
- 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
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
VirusTotal
64/64 vendors flagged this skill as clean.