Security audit
deepresearchpro
Security checks across malware telemetry and agentic risk
Overview
This is a transparent deep-research skill that guides web-based source gathering and report writing, with no evidence of hidden credential access, persistence, destructive behavior, or data exfiltration.
Install if you want an agent-guided deep research workflow and are comfortable with your queries being sent to external search engines and websites. Review final citations and publication dates carefully, especially when a source page does not clearly state a date.
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
55/55 vendors flagged this skill as clean.