ClawHub
Back to skill

Security audit

Claude代码运行器

Security checks across malware telemetry and agentic risk

Overview

This skill is a coding-agent runner with expected command and file access, but it asks for broad execution/write authority and mentions root/sudo user switching without tight scoping or confirmation guidance.

Install only if you intend to let the agent run commands and edit code through a PTY. Avoid using it with sudo/root unless you have isolated the environment, and prefer a disposable workspace or least-privilege user for repository changes and deployment-related tasks.

SkillSpector

Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
84% confidence
Finding
The skill advertises very broad activation conditions such as code generation, debugging, testing, and deployment, without a precise invocation boundary. In an agent setting, this can cause the skill to be invoked in overly many contexts and route sensitive tasks into a toolchain that includes execution and file-write capability, increasing the chance of unsafe or unintended actions.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill promotes code execution, write access, and deployment-oriented usage but does not prominently warn users that it may execute commands or change files. In this context, the available tools include exec and write, so weak disclosure materially increases the risk that users or orchestrating agents trigger impactful operations without informed consent.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Content
配置后需重启会话或开启新终端生效。API Key应妥善保管,避免泄露到版本控制系统.
## 限制条件
* Requires Unix-like environment (uses PTY)
* Requires root/sudo for user switching
* ai-assistant Code must be installed separately
* May not handle all edge cases of interactive prompts
Confidence
87% confidence
Finding
The documentation states that root/sudo may be required for user switching, which elevates the risk profile of this skill substantially. In a skill explicitly designed to run coding tasks through a PTY with exec capability, encouraging privileged execution can amplify any prompt mistake, command injection, or unintended action into full system compromise or destructive host changes.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.