Security audit
小五子 - 基因修复
Security checks across malware telemetry and agentic risk
Overview
The skill is instruction-only and not overtly harmful, but it directs automatic rollback and prompt/instruction repair for agents without clear limits or user approval.
Review before installing in any agent that can modify its own prompts, skills, or long-running task state. Use it only with explicit approval for repairs or rollbacks, clear logs, and boundaries on what agent state it may change.
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
63/63 vendors flagged this skill as clean.