Security audit

AI大模型专家｜GEO+AEO 品牌 AI搜索增长

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed AI-HIVE planning and media-generation helper, with credential and upload behavior that users should understand before use.

Install only if you intend to use AI-HIVE for planning and media generation. Treat the API key as a real credential, prefer environment variables or a local 0600 config file, and upload only assets you are authorized to send to the service. Review generation costs and routing before submitting tasks.

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 (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 89% confidence
Finding: The skill advertises and demonstrates code paths that use environment variables, local file I/O, network access, and shell execution, yet it does not declare permissions. This creates a trust and sandboxing gap: users and the hosting platform cannot accurately assess what the skill can access, increasing the chance of over-broad execution and accidental exposure of local data or secrets.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 95% confidence
Finding: The documented purpose is GEO/AEO planning and media generation orchestration, but the behavior reportedly extends to user/account queries, wallet balance access, arbitrary model enumeration, generic chat calls, and browser-assisted API key capture and local storage. This mismatch is dangerous because users may grant trust for a narrowly scoped content skill while it performs broader account, credential, and platform actions that could expose billing data, secrets, or enable unintended API usage.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The implementation materially diverges from the declared skill purpose: instead of a GEO/AEO brand-search growth expert, it provides a generic AI-Hive client with image/video generation, uploads, task polling, and account-related operations. This kind of scope mismatch is dangerous because it can smuggle undeclared capabilities past reviewers and grant users or downstream agents access to media-generation and platform API actions they would not reasonably expect from the manifest.

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The file exposes extra capabilities such as user-info retrieval, model enumeration, API-key setup, and browser launching that are unrelated to the stated consulting/planning use case. In the context of a mislabeled skill, these broaden the attack surface and can enable unintended account inspection, service discovery, or user-environment interaction beyond what a user would expect.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The skill enables implicit invocation without any trigger constraints, which allows the agent to select and run this skill in broader contexts than the user may have intended. Because this skill is positioned to drive planning and executable generation workflows, unintended invocation could cause prompt-scope expansion, irrelevant data collection, or unauthorized task execution based on ambiguous user input.

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: SKILL.md:133