Research Harness v1.3
Security checks across malware telemetry and agentic risk
Overview
This is a documentation-only research workflow skill that gives agents governance and reproducibility guidance without hidden code, persistence, or credential access.
Install if you want agents to follow a stricter research workflow. Expect it to influence how agents organize experiment repos, write validation scripts, preserve failed runs, and request human review before changing research claims, rubrics, or schemas.
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.