Aibrary Podcast Dialogue
Security checks across malware telemetry and agentic risk
Overview
This is a simple instruction-only writing skill for generating two-person podcast scripts about books, with no evidence of unsafe access or hidden behavior.
This skill appears safe to install from the reviewed artifacts. Users should still verify factual claims in generated book discussions and avoid providing private or unpublished book content unless they are comfortable using it as prompt input.
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
66/66 vendors flagged this skill as clean.