ClawHub

code-with-claude

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only Claude Code reference skill; it describes powerful features but does not install or run them itself.

Safe to install as reference material, but do not blindly copy examples involving hooks, HTTP endpoints, agent hooks, remote control, permission bypassing, or long-lived tokens. Review what data a hook or channel may send, prefer trusted local or HTTPS endpoints, and keep tool permissions narrowly scoped.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (2)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The documentation explicitly describes an `http` hook type that automatically POSTs event JSON to a URL, but it does not warn that hook events may contain sensitive prompts, tool metadata, file paths, or other session context that could be transmitted off-host. In a CLI/plugin context where hooks fire automatically at many lifecycle events, this omission can lead users to enable silent data exfiltration without understanding the privacy and security implications.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The `prompt` and especially `agent` hook types are documented as automatically executing at lifecycle events, with the agent variant having full tool access, but the text does not warn about scope, autonomous execution, or side effects. In this skill's context—a reference for configuring Claude Code hooks—users may copy these patterns directly, creating automatic LLM- or tool-driven actions that can inspect data, modify files, invoke commands, or contact external systems with insufficient oversight.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal