ClawHub
Back to skill

Security audit

CertainLogic Self-Improving Stack

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only skill bundle that recommends installing memory and self-improvement tools, with no hidden code or automatic execution in the artifact.

Before installing, review each referenced skill individually, especially the memory, learning, proactive, hermes-learning-loop, and Company Brain tools. Confirm where they store data, whether they share information across tools, and how to disable or delete persisted memory. Treat the vendor verification claims as marketing context, not a substitute for reviewing the dependent skills.

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
Findings (1)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The document explicitly describes storing corrections, user-specific data, history, and extracted skills across multiple tools without a clear privacy notice, retention policy, or consent guidance. In a self-improving agent context, persistent storage of user interactions can capture sensitive prompts, personal data, secrets, or business information and make later leakage or unintended reuse more likely.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.