ClawHub
Back to skill

Security audit

Model Usage

Security checks across malware telemetry and agentic risk

Overview

This skill locally summarizes CodexBar cost logs and does not show hidden data transfer or unrelated system access.

Install only if you are comfortable adding the CodexBar CLI via Homebrew and allowing it to read local Codex or Claude usage logs for cost reporting. Review CodexBar itself if you have strict privacy requirements for local session logs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
Findings (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
85% confidence
Finding
The skill instructs the agent to invoke shell commands (`codexbar`, `python`) and read local files/stdin (`--input /tmp/cost.json`, `cat /tmp/cost.json`) but does not declare corresponding permissions. This creates a trust and enforcement gap: a host may treat the skill as low-privilege while it actually requires filesystem and shell access, increasing the chance of unintended command execution or local data exposure.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.