ClawHub
Back to skill

Security audit

subagent-orchestrator

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only orchestration skill that openly coordinates subagents and persistent task records, with no executable install code or hidden data access.

Install this only if you want OpenClaw to run multi-step subagent workflows with less repeated prompting. Keep task boards explicit about goals, stop conditions, and allowed scope; review MEMORY.md and daily memory records for sensitive or stale instructions; and preserve screenshots or proof artifacts before cleanup when they may be needed.

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 (1)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The invocation description is broad enough to match ordinary project-management or workflow language, which can cause this skill to activate in situations beyond true subagent orchestration. Because the skill then instructs the controller to continue dispatching work automatically and not wait for further user confirmation, over-triggering can lead to unintended autonomous actions, task continuation, and state changes.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.