ClawHub
Back to skill

Security audit

Adobe Firefly 图片生成替代｜AI 图片生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI image generation/editing wrapper that sends user-chosen images and prompts to AI Hive and stores an API key for later use.

Install only if you are comfortable sending selected prompts and reference images to AI Hive. Use non-sensitive images unless you trust the provider, protect the AI Hive API key, and review returned/downloaded files like any other remote content.

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
95% confidence
Finding
The skill invokes local Python scripts, installs dependencies, reads input images, may write outputs, accesses environment variables via shell execution context, and sends data to a remote API, yet it declares no permissions. This creates a mismatch between the skill's documented trust boundary and its actual capabilities, preventing users or the platform from making informed security decisions about network exfiltration, file access, and command execution.

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The code claims requests always go only to the fixed AI Hive address, but it also performs uploads to arbitrary upload URLs returned by the API and downloads arbitrary result URLs from task responses. If the API, an intermediate service, or returned metadata is compromised, the client can be induced to connect to attacker-controlled hosts, enabling SSRF-like outbound access, untrusted content retrieval, or misleading trust assumptions by users.

Static analysis

No suspicious patterns detected.