Security audit

美妆步骤教程短视频｜AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-HIVE video-production helper that uses user-provided media and credentials for expected generation workflows.

Install only if you intend to use AI-HIVE for video generation and are comfortable providing an AI-HIVE API key, uploading authorized reference media, and potentially incurring generation costs. Review generated commands and parameters before running them, and use an environment variable instead of saved config if you do not want the API key persisted locally.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (6)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 90% confidence
Finding: The skill embeds executable workflows that use environment variables, filesystem access, shell commands, and network calls, but it does not declare corresponding permissions. This creates a transparency and policy-enforcement gap: a host may invoke the skill without understanding it can read/write local files, access API keys, upload media, or call external services, increasing the chance of unintended data exposure or unauthorized actions.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 95% confidence
Finding: The documented purpose is narrowly framed as beauty tutorial short-video generation, but the described behavior includes generic chat, image generation, media upload, account/model querying, browser-based API-key setup, and broad ffmpeg processing. That mismatch makes the skill more dangerous because users and policy systems may trust it as a domain-specific helper while it actually exposes general-purpose content-generation and local-system capabilities that could be repurposed for unrelated or risky operations.

Description-Behavior Mismatch

Medium

Confidence: 97% confidence
Finding: The file is packaged and described as a narrow beauty-tutorial video skill, but it contains a broad general-purpose OpenAPI client with chat, model enumeration, account inspection, and arbitrary media upload primitives. This capability mismatch expands the trust boundary: a caller expecting a constrained video skill may unknowingly grant access to broader account and platform operations than necessary.

Context-Inappropriate Capability

Low

Confidence: 89% confidence
Finding: The skill exposes a user-info command that retrieves account information and wallet balance, which is unrelated to generating beauty tutorial videos. While not inherently code-execution dangerous, it unnecessarily grants access to sensitive account metadata and can leak billing or identity information in a context where users would not expect it.

Context-Inappropriate Capability

Low

Confidence: 87% confidence
Finding: The model-listing capability lets the skill enumerate the broader provider model inventory, which is outside the stated beauty-video purpose. This increases reconnaissance value for an attacker or overly powerful agent by revealing available platform capabilities that could later be abused through the same generic client.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: Enabling implicit invocation without any activation constraints can cause the skill to trigger in situations the user did not explicitly intend, increasing the chance that its default workflow and code-generation behavior are applied automatically. In a skill that can produce runnable AI-HIVE commands and media-generation workflows, this raises the risk of unintended tool use, unexpected costs, or processing of sensitive commercial inputs without clear user confirmation.

Static analysis

No suspicious patterns detected.