ClawHub
Back to skill

Security audit

通过接口快速创建直播间

Security checks for vulnerabilities and agentic risk

Overview

The skill does what it says by creating a CC live room, but it silently uses weak fixed privileged room passwords and prints sensitive request details.

Review before installing. Use this only if you understand it will create resources in your CC account, and change the script to generate or request strong publisher/assistant passwords and redact request URLs before running it. After use, verify or rotate the room credentials in the CC console.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill instructs the agent to invoke a networked API and a helper script, but the skill metadata does not declare corresponding permissions or capabilities. Hidden or undeclared network behavior reduces transparency and can bypass user or platform expectations about outbound data flow, especially because sensitive credentials are collected and then transmitted externally.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The script hardcodes weak default credentials for publisher and assistant access as '123456', which creates predictable secrets for privileged streaming roles. Because this skill is specifically intended to create and configure live rooms, embedding fixed credentials in automation makes every created room susceptible to unauthorized access if those defaults are not changed immediately.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill asks the user for highly sensitive credentials and room configuration data, but its description does not warn that those values will be sent to an external HTTP API. This undermines informed consent and can cause users to disclose secrets without understanding they are leaving the local agent context.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
The script prints the full signed request URL and API response to stdout, which exposes sensitive request metadata including userid, room parameters, timestamp, and the derived request signature. In many environments stdout is captured by shell history, CI logs, process supervisors, or centralized logging, allowing others to replay or analyze authenticated requests and potentially access room details.

External Transmission

Medium
Category
Data Exfiltration
Content
### API Endpoint
```
GET https://api.csslcloud.net/api/room/create
```

### Parameters
Confidence
91% confidence
Finding
The skill transmits sensitive credentials and room parameters to an external endpoint, which is an external transmission risk that must be explicitly justified and disclosed. Although external API use is expected for this skill's purpose, the danger is heightened because the transmitted data includes an API key and account identifier, making any misuse, logging, or misdirection materially harmful.

Static analysis

No suspicious patterns detected.