Multi-Agent Debate
Security checks across malware telemetry and agentic risk
Overview
This is an instruction-only skill for structured multi-agent debate, with no code, installs, credential use, or hidden system access.
This skill appears safe to install as a prompt workflow. Expect higher model cost and latency because it asks for multiple independent answers and critique rounds. Avoid placing secrets or highly sensitive information into debate prompts unless your agent environment is approved for that data, since the same content may be processed across multiple internal reasoning passes.
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.