Security audit

AI漫剧全流程制作｜AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-HIVE media-production helper, and I did not find hidden exfiltration, destructive behavior, or automatic high-impact actions.

Install only if you are comfortable using AI-HIVE for media generation, uploading the reference files you choose, and optionally storing an API key locally. Review prompts, routing mode, uploaded media, and likely cost before running generate commands, especially for batch jobs.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (4)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 90% confidence
Finding: The skill advertises executable code paths that can access environment variables, read/write local files, invoke shell commands, and make network requests, yet it declares no permissions or capability boundaries. This is dangerous because users and host systems cannot make an informed trust decision, and the combination of shell, filesystem, network, and env access could expose secrets or enable unintended system-side actions if the skill is run in an agent environment.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 87% confidence
Finding: The documented purpose focuses on comic-drama production workflow, but the described behavior extends to generic text chat, account/model enumeration, media upload utilities, and broad local ffmpeg editing operations. This mismatch is dangerous because it hides materially broader capabilities than users would reasonably expect, increasing the risk of covert data handling, unauthorized uploads, or local file manipulation under the cover of a narrowly described creative workflow.

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The file implements a broad-purpose AI Hive CLI with chat, image generation, model enumeration, user-info lookup, uploads, and task polling, while the skill metadata presents a much narrower comic-drama video workflow. This capability mismatch increases attack surface and enables use of the skill for actions not expected or approved by users or platform policy, weakening least-privilege and making misuse easier.

Context-Inappropriate Capability

Low

Confidence: 84% confidence
Finding: The skill exposes a user-info endpoint that retrieves account information and wallet balance even though the stated purpose is content-production workflow automation. Unnecessary access to account metadata expands the data surface and may leak sensitive billing or identity details if invoked by an agent or user unexpectedly.

Static analysis

No suspicious patterns detected.