Security audit
Weclone Twin Reply
Security checks across malware telemetry and agentic risk
Overview
This skill builds a review-only draft reply from user-provided persona and chat files, with no evidence of hidden sending, exfiltration, or destructive behavior.
Install only if you are comfortable with selected persona files and recent conversation snippets being assembled into a model prompt. Keep the approval gate, do not auto-send drafts, avoid storing secrets in ai_twin/, and remove unrelated Markdown files from that directory before rendering.
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
65/65 vendors flagged this skill as clean.
Static analysis
No suspicious patterns detected.