ClawHub
Back to skill

Security audit

my-tools

Security checks across malware telemetry and agentic risk

Overview

This is a simple time and calculator skill, but its calculator can run arbitrary Python expressions instead of only doing math.

Review carefully before installing. The skill does not show hidden theft, persistence, or destructive behavior, but the calculator should be considered unsafe until it replaces eval() with a strict arithmetic parser and narrows when it runs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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 (3)

eval() call detected

High
Category
Dangerous Code Execution
Content
import sys
try:
    expr = sys.argv[1]
    result = eval(expr, {"builtins": {}}, {})
    print(f"{expr} = {result}")
except Exception as e:
    print(f"计算失败: {str(e)}")
Confidence
96% confidence
Finding
result = eval(expr, {"builtins": {}}, {})

Vague Triggers

Medium
Confidence
82% confidence
Finding
The time skill is triggered by very broad keywords such as '时间', '日期', and '几点', which can appear in ordinary conversation without a real request to invoke a tool. That can cause unnecessary execution of the exec tool and make routing unreliable, increasing the chance of unintended tool use in contexts where the user was not actually asking for current time/date data.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The math trigger conditions are overly broad and the skill requires passing the user's expression to calc.py via exec, creating a larger attack surface for command or argument injection depending on how downstream tooling handles quoting and argument parsing. Because even loosely math-like text may activate the skill, an attacker has more opportunities to smuggle unexpected input into tool execution.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

Detected: suspicious.dynamic_code_execution

Dynamic code execution detected.

Critical
Code
suspicious.dynamic_code_execution
Location
calc.py:4