ClawHub
Back to skill

Security audit

Nano Banana Pro 文生图

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent AI Hive text-to-image helper that discloses its API key setup, network use, task polling, and image downloads.

Before installing, be aware that this skill sends your prompts to AI Hive, stores an AI Hive API key locally if you run init, and downloads generated files to a local output directory. Use a dedicated API key if possible and avoid submitting private or sensitive prompt content unless you trust the service.

SkillSpector

Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill advertises executable capabilities including shell, network, environment access, and local file read/write, yet no permissions are explicitly declared. This creates an authorization transparency gap: users and hosting platforms cannot accurately assess or constrain what the skill may do, increasing the chance of unintended secret access, local file modification, or network misuse.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The description presents a simple text-to-image skill, but the documented behavior includes local credential setup, persistent key storage, arbitrary task lookup by task ID, and filesystem downloads. This mismatch is dangerous because users may invoke the skill without realizing it handles secrets and local state, and arbitrary task querying can expose or retrieve generation metadata or outputs beyond the immediate user workflow if task IDs are guessable or shared.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.