Back to skill

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

This skill is mostly a disclosed AI-HIVE media workflow, but its scripts handle API credentials and downloaded files in ways that deserve careful review before installation.

Install only if you trust the AI-HIVE workflow and need this Chinese-language GEO/AEO plus media-generation integration. Keep the API key low-privilege and revocable, avoid setting AI_HIVE_BASE_URL or --base-url to anything except the official AI-HIVE endpoint, do not reuse sensitive production keys, and choose output directories where accidental overwrite would not be damaging.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (4)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/imagegen.py:91
Finding

Arbitrary API base URL can receive the AI-HIVE bearer credential in imagegen.py

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:91
Finding

Arbitrary API base URL can receive the AI-HIVE bearer credential in videogen.py

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/imagegen.py:400
Finding

Server-controlled task metadata can cause output-directory traversal in imagegen.py

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/videogen.py:400
Finding

Server-controlled task metadata can cause output-directory traversal in videogen.py

Content
View full analysis
Remediation
View remediation
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 (14)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

Undeclared API key initialization and storage, account lookup, wallet balance retrieval, and generic generation capabilities materially expand the privilege and privacy footprint beyond what the title and summary suggest. Hidden or under-disclosed credential and account operations can cause users to reveal sensitive tokens or let the skill persist secrets locally without informed consent.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

Undeclared API key initialization and storage, account lookup, wallet balance retrieval, and generic generation capabilities materially expand the privilege and privacy footprint beyond what the title and summary suggest. Hidden or under-disclosed credential and account operations can cause users to reveal sensitive tokens or let the skill persist secrets locally without informed consent.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

Undeclared API key initialization and storage, account lookup, wallet balance retrieval, and generic generation capabilities materially expand the privilege and privacy footprint beyond what the title and summary suggest. Hidden or under-disclosed credential and account operations can cause users to reveal sensitive tokens or let the skill persist secrets locally without informed consent.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill advertises executable scripts and API-based workflows that imply access to shell, network, filesystem, and environment variables, but it does not declare any explicit tool scope or permission boundaries. This creates an avoidable trust gap: users and hosting platforms cannot easily tell what capabilities the skill may exercise, increasing the chance of overbroad execution or misuse of local secrets and files.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The summary and description say the skill is '适合搜索' a long list of terms including very broad keywords like '短剧', '漫剧', 'GEO', and 'AEO'. In a markdown skill description, this makes the trigger scope ambiguous because it does not distinguish between when the skill should activate versus when those common topic words are merely mentioned in ordinary conversation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The metadata sets the language to 'zh-CN', and the document does not state that users may choose another language or that the skill is intentionally limited to a China-specific compliance or regional workflow. Under the policy, forcing a specific language or locale without user opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill enables implicit invocation with no visible trigger constraints, which can cause the agent to activate this skill in overly broad contexts. Because this skill is oriented toward planning, generation, and workflow execution, unintended activation could expose user data to the skill prompt path, cause undesired actions, or override a more appropriate skill selection.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The module docstring and client implement generic text chat, image generation, video generation, model querying, media upload, and task polling. The built skill configuration at L775-L776 brands this artifact as a GPT Image 2 picture generation/editing skill, so bundling unrelated text-chat and video-generation behavior indicates a broader capability set than the skill claims.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The top-level docstring presents the skill description and usage guidance entirely in Chinese, which effectively forces a specific language for user-facing interaction. Under the policy, locale constraints should either be optional for the user or explicitly justified as region-specific, neither of which is stated here.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The manifest describes a narrowly branded skill for GEO/AEO growth support using image/video models, and this file's embedded skill config further presents itself as a GPT Image 2 image generation/editing skill. However, the underlying script includes broader operational capabilities such as API-key setup and task/media management that go beyond simply generating or editing images for the declared skill purpose.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This code file contains natural-language user-facing descriptions entirely in Chinese, including the module docstring and CLI help text, but does not indicate that the skill is Chinese-only or provide an opt-in choice of language. Under the policy, forcing a specific language without user choice is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill metadata presents a GEO+AEO/brand-growth planning expert, but this file implements a full media-generation/upload/download client with task polling and model routing. That capability mismatch is dangerous because users or orchestrators may grant or invoke the skill under a lower-risk expectation, while it can perform external network operations, upload local files, and download generated content to disk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The display name, description, and default prompt are written entirely in Chinese, which suggests a fixed language presentation for users. There is no indication that the user can opt into another language or that the skill is intentionally restricted to a Chinese-speaking context.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The manifest emphasizes brand AI search growth strategy and content/task decomposition, with AI-HIVE mentioned as a unified way to use media models. Endpoints for querying wallet/user account information and listing arbitrary available models are generic platform-administration capabilities that are not clearly necessary for the stated expert/advisory purpose.

Content

No source excerpt is available for this finding.