Back to skill

Security audit

商品图转UGC带货视频|AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed AI-HIVE media-generation workflow, but it needs review because it stores and uses an API key and can send that key to an arbitrary configured API endpoint.

Install only if you trust the AI-HIVE account integration and are comfortable storing an API key locally. Do not set or accept custom AI_HIVE_BASE_URL/base-url values unless you control the endpoint, use a limited API key, review commands before running them, and expect uploaded product media to be sent to the external AI-HIVE service.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/imagegen.py:91
Finding

Bearer Credential Disclosure Through an Unrestricted API Base URL

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
scripts/imagegen.py:7
Finding

Unpinned Third-Party Dependency Installation

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (17)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The skill claims a narrow, product-image-to-UGC-video purpose but reportedly includes undeclared capabilities such as generic chat, account or wallet retrieval, and model discovery while failing to provide several promised deliverables. Undeclared features expand the attack surface and can surprise users with access to account-related data or unrelated functions that were not consented to or reviewed in the skill description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill claims a narrow, product-image-to-UGC-video purpose but reportedly includes undeclared capabilities such as generic chat, account or wallet retrieval, and model discovery while failing to provide several promised deliverables. Undeclared features expand the attack surface and can surprise users with access to account-related data or unrelated functions that were not consented to or reviewed in the skill description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill claims a narrow, product-image-to-UGC-video purpose but reportedly includes undeclared capabilities such as generic chat, account or wallet retrieval, and model discovery while failing to provide several promised deliverables. Undeclared features expand the attack surface and can surprise users with access to account-related data or unrelated functions that were not consented to or reviewed in the skill description.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The packaged skill advertises product-image-to-UGC-video generation, but its exposed entrypoint only wires the generate command to cmd_image, not cmd_video. This capability mismatch is dangerous because agents and users may trust the skill to perform one bounded media task while it actually exposes different behavior, undermining policy gating, auditability, and informed user consent about what will be produced.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill advertises and demonstrates capabilities that require environment access, filesystem operations, shell execution, and network calls, but it does not declare any explicit tool scope or permission boundaries. In an agent environment, this increases the risk of over-privileged execution, accidental secret exposure, unintended file access, or unauthorized outbound requests because neither users nor the platform can enforce least privilege from the manifest.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The skill content, triggers, headings, examples, and expected outputs are entirely written in Chinese, with no statement that users may choose another language. This can violate a language/locale policy when a skill implicitly requires a specific language without opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The default prompt hard-codes Chinese output and a specific workflow shape regardless of the user's language preference. While not a direct security flaw on its own, it can override user intent and reduce transparency, which is riskier here because the skill can drive external service operations and generated marketing content that the user may not fully understand or review.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill enables implicit invocation while its activation scope is described broadly in metadata and centered on marketing/e-commerce use cases. This increases the chance the agent auto-triggers the skill in adjacent contexts without clear user intent, which can cause unintended API actions, content generation, or uploads tied to a third-party service.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

This code hard-codes the skill description, skill name, and search aliases in Chinese, including the CLI parser description and generated metadata. Under the policy rule, forcing a specific language or locale without offering a user choice can be a natural-language policy violation, and no opt-in or documented locale justification appears in this file.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/edit_video.py (reported line 20)May include surrounding context.

python
def run(args: list[str]) -> None:
    print(" ".join(args))
    subprocess.run(args, check=True)


def main() -> None:

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring describes the code as a general AI capability invocation tool covering chat, image, video, model query, upload, and polling, while the packaged skill claims a narrow specialized purpose. This inconsistency is a supply-chain transparency problem: reviewers and orchestrators may approve the skill under false assumptions, while the codebase actually carries broader operational capability.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

This code file contains its primary docstring and most user-facing CLI messages in Chinese, including setup, error, and help text, but does not offer any language selection or indicate that the skill is intentionally China-region-specific. That creates a natural-language policy issue under the locale/language rule because the skill implicitly enforces one language for all users without opt-in.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The file contains a general-purpose chat subcommand that accepts arbitrary prompts and optional images, which exceeds the narrowly declared product-image-to-UGC-video skill scope. In a skill ecosystem, this creates unnecessary latent capability that can be invoked directly or repurposed to bypass least-privilege expectations, increasing prompt-injection and policy-scope risk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The manifest describes a narrowly scoped skill for turning product images into UGC/e-commerce videos, but this file exposes generic chat, image generation, user-info, model listing, arbitrary media upload, and general task-query capabilities. Those operations are not merely hidden implementation details of the declared workflow; they make the script usable as a general-purpose AI-Hive API tool.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The top-level docstring and parser description explicitly label the program as a '通用 AI 能力调用工具' and enumerate text chat, image generation, model query, media upload, and polling features. That directly conflicts with the specialized manifest for a product-image-to-UGC-video skill, so the in-code documentation presents a materially different intent than the actual skill packaging claims.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

This Python file contains user-facing descriptions, help text, status messages, and setup instructions almost entirely in Chinese, including the module docstring and CLI interaction text. Because the file does not offer a language choice or document a justified region-specific restriction, it violates the language/locale policy for natural-language content.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The user-info endpoint retrieves user information and wallet balance, and the CLI exposes it directly via cmd_user_info. Accessing account profile/balance data is not justified by the manifest’s stated purpose of transforming product images into marketing videos, and it introduces a separate capability unrelated to content generation.

Content

No source excerpt is available for this finding.