ClawHub

Ilya.skill - Ilya Sutskever思维操作系统

Security checks across malware telemetry and agentic risk

Overview

This is a conversational persona skill with no executable code or system access, but users should treat it as simulation rather than a real person’s advice.

Install only if you want a simulated Ilya Sutskever-inspired reasoning style. Treat outputs as roleplay or an analytical lens based on public material, not as authentic statements, endorsement, or expert advice from Ilya Sutskever, especially for AI safety, business strategy, investment, or other high-stakes decisions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest frames the skill as providing analysis from Ilya Sutskever's perspective, but the body explicitly instructs the agent to impersonate him directly and to minimize repeated disclosure. This creates a deceptive identity/policy mismatch that can mislead users about authorship, provenance, and accountability, especially in safety-sensitive discussions where users may over-trust the apparent authority of the persona.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger conditions are broad enough to activate on ordinary conversational references like 'Ilya would think' or 'switch to Ilya,' which can cause unintended persona takeover without clear user consent. In practice this can override normal assistant behavior and inject impersonation-style responses into benign queries, increasing the chance of confusion, misrepresentation, or unsafe reliance on stylized authority.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal