emergency circuit

Security checks across malware telemetry and agentic risk

Overview

The skill is a disclosed AI-agent monitoring and kill-switch guide, with high-impact shutdown behavior that is expected for its purpose but should be used carefully.

Before installing, verify the external npm or GitHub implementation, test policies in sandbox mode, use explicit agent IDs, and confirm how logs and incident data are stored. Treat kill and auto-kill policies as production-impacting controls that may stop active workflows or leave external actions partially complete.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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

Findings (2)

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The documentation repeatedly promotes automatic shutdown and emergency termination of running agents, but it does not clearly warn users that triggering the kill switch may immediately interrupt live workflows, drop in-flight tasks, or leave external side effects partially completed. In a production monitoring skill, this omission can cause operators to use the feature without understanding the operational and data-consistency risks.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The documented `emergency-circuit kill --agent-id <id>` command is presented as a normal operational command without an explicit caution that it performs an immediate destructive stop. Users may invoke it in production without realizing it can abruptly terminate active processing and create service interruption or inconsistent downstream state.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal