Security audit
Swarm Orchestrator
Security checks across malware telemetry and agentic risk
Overview
This skill is a disclosed local multi-agent orchestration setup with expected local services, optional LLM API keys, and no artifact evidence of hidden data collection or destructive behavior.
Install only after reviewing the external npm package or GitHub repo, prefer Docker or another isolated environment, keep ports bound to localhost, add LLM API keys only when needed, and require human approval for tool/custom agents that can run scripts, modify files, access accounts, or call external systems.
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
66/66 vendors flagged this skill as clean.
Static analysis
No suspicious patterns detected.