Requirements Traceability TraceMap
Security checks across malware telemetry and agentic risk
Overview
This is a straightforward traceability-reporting skill with no hidden execution, persistence, credential handling, or privileged behavior in the submitted artifacts.
Safe to install for requirements traceability and QA reporting. Because the skill may analyze PRDs, code diffs, configs, prompts, model changes, and test evidence, only provide repositories or documents you intend the agent to review, and treat its release recommendation as advisory rather than an automatic approval gate.
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
53/53 vendors flagged this skill as clean.