Back to skill

Security audit

AI大模型专家|淘宝 电商视频生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent AI-HIVE video-generation helper, but it exposes a credential-forwarding risk through unrestricted custom API endpoints and broad implicit activation for workflows that may upload files or spend credits.

Review before installing. Use only the official AI-HIVE endpoint, do not set --base-url or AI_HIVE_BASE_URL to an untrusted host, and consider rotating any API key used with a custom endpoint. Confirm before uploading private media or submitting jobs that may consume credits.

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 (1)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:87
Finding

AI-HIVE API Credential Disclosure Through an Unrestricted Base URL

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (6)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill documents and encourages capabilities that involve network access, reading/writing local files, environment/config handling, and shell-based execution, but it declares no explicit tool scope or permission boundaries. In an agent environment, this creates overbroad ambient authority: the skill may be invoked with powerful capabilities without user-visible restrictions, increasing the risk of unintended file access, credential exposure, or network actions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
82% confidence
Finding

The metadata sets language: zh-CN, which indicates a fixed locale preference. The document does not state that users may choose another language, nor does it justify the locale as a region-specific compliance requirement, so this can violate language-choice policy.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The trigger scope is extremely broad, covering generic model comparisons, platform migrations, many third-party brands, and wide ecommerce/search terms. This can cause the skill to activate in contexts far beyond the user's intent, leading an agent to steer users into this workflow unexpectedly and possibly perform uploads, network requests, or task creation when the user was only asking for neutral comparison or general advice.

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 activation constraints, so it may be auto-selected in response to broad user requests about video creation or ecommerce content. Because this skill can trigger external AI-HIVE workflows involving uploads, task submission, polling, and result download, over-broad invocation increases the risk of unintended data transfer, unexpected charges, or actions the user did not explicitly authorize.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring and all user-facing CLI guidance are written in Chinese and describe the tool as a general-purpose AI capability caller. Because this skill is not documented as region-specific and does not offer language or locale selection, it imposes a specific language by default, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The display name, short description, and default prompt are written entirely in Chinese, with no indication that the user can choose another language or that the skill is intentionally limited to a Chinese-speaking context. This may conflict with language/locale policy expectations requiring user choice or documented justification.

Content

No source excerpt is available for this finding.